About SAFLII
Databases
Search
Terms of Use
RSS Feeds
South Africa: Competition Appeal Court
SAFLII
>>
Databases
>>
South Africa: Competition Appeal Court
>>
2015
>>
[2015] ZACAC 4
|

|

Sasol Chemical Industries Limited v Competition Commission (131/CAC/Jun14) [2015] ZACAC 4; 2015 (5) SA 471 (CAC); [2015] 1 CPLR 58 (CAC) (17 June 2015)

REPUBLIC
OF SOUTH AFRICA
IN
THE COMPETITION APPEAL COURT OF SOUTH AFRICA
HELD IN CAPE TOWN
CASE
NO: 131/CAC/Jun14
DATE:
17 JUNE 2015
Reportable
In
the matter between:-
SASOL
CHEMICAL INDUSTRIES
LIMITED
....................................................................
Appellant
And
THE
COMPETITION
COMMISSION
...............................................................................
Respondent
JUDGMENT:
17 June 2015
DAVIS
JP
Introduction
[1]
This case concerns the meaning and the
scope of s 8(a) of the Competition Act 89 of 1998 (‘the Act’)
which provides
that a dominant firm may not ‘charge an
excessive price to the detriment of consumers. Section 1 (1) defines
an “excessive
price” as a price for a good or service
which (aa) bears no reasonable relation to the economic value of that
good or service;
and (bb) is higher than the value referred to in sub
paragraph (aa).’
[2]
While economists define excessive prices as
those which are significantly and persistently above the competitive
level, the translation
of this seemingly simple statement into law
and consequently the enforcement of s 8(a) of the Act in immensely
complex.
David Lewis (
Thieves
at the Dinner Table
( 2012) at 177
notes that excessive pricing in competition law is fraught with
complexity and controversy:
‘
Pricing
power derives from market power.  However, the mere possession
of market power is not contrary to competition law.
Indeed some
important source of market power is innovation and other,
pro-competitive conduct.  The rents derived from the
possession
of market power will, in most circumstances, sooner or later attract
new entrants, the more so if the dominant incumbent
takes ‘excessive’
advantage of its privileged position.  And so the effort to
acquire market and, therefore, pricing
power and the attention it
attracts from rivals are an important driver of the competitive
process.’
A similar sentiment
is articulated by O’Donoghue and Padilla
The Law and
Economics of Article 102 TFEU
(2
nd
ed) at 737:
‘
The
use of Article 102 TFEU to curb excessive prices has been criticised
on several different levels.  A first basic objection
is that no
generally accepted criterion exists in the decisional practice and
case law to determine when price are “excessive.”

Further, even if a criterion, or series of criteria, could be agreed
upon as a benchmark, determining an excessive price in practice
is
extremely complex and subject to a number of difficulties.  A
second criticism is that prices above marginal cost are common
and
necessary in many industries where high profits are necessary to
recover large up-front capital and other fixed costs.
Third,
any policy on excessive prices is likely to yield incorrect
predictions in many instances and the costs of such errors is
likely
to be higher than the costs of allowing certain excessive prices to
escape censure.  Fourth little or no guidance is
offered in the
decisional practise and case law on what might constitute objective
justification for a price that exceeds the relevant
criterion for
determining an excessive price.  Finally, devising effective
remedies in excessive pricing cases raises difficult
issues.’
This
case compels this court to engage with all of this complexity and
controversy which is inherent in the doctrine of excessive
pricing.
The
Factual Matrix
[3]
The case involves an appeal by appellant
against the decision of the Competition Tribunal (‘Tribunal’)
that it contravened
s 8 (a) of the Act by charging excessive prices
for propylene and polypropylene (PP) from 2004 to 2007.  The
production of
purified propylene is dependent on the production of
purified feedstock propylene.  Polypropylene is produced from
purified
propylene.   Appellant is the only significant
producer of purified propylene in South Africa.
[4]
In August 2010 respondent (‘the
Commission’) referred three complaints against appellant and
Safripol (Pty) Ltd to the
Tribunal submitting that both companies had
contravened the Act.  One of the complaints was a price fixing
complaint.
This was settled in 2010.  The two
remaining complaints concern forms of excessive pricing which are the
subject matter of
this appeal.
[5]
In essence, the Commission contends that
appellant charged excessive prices for PP within the period 2004 to
2007.  During
this period appellant was dominant in the market
for the supply of PP to domestic customers in South Africa, having a
market share
between 64% and 80% over the complaint period.
According to the founding affidavit deposed to by Mr Lesofe on behalf
of the
Commission:
‘
Up
until 2008, appellant had the capacity to produce approximately
228 000 tonnes of PP per year.  As from January 2008,

Sasol’s capacity has expanded to approximately 528 000
tonnes per year.
Safripol
(Pty) Ltd has the capacity to produce approximately 120 000
tonnes of PP per year.
Both
appellant and Safripol sell PP in South Africa, primarily to plastic
product manufacturers.  However, appellant and Safripol
together
produce more polypropolene than is required by customers in South
Africa.  Domestic demand for PP has remained constant
since 2004
at approximately 220 000 tonnes per year.
Consequently,
appellant (and to a lesser extent Safripol) export significant
quantities of polypropylene both to other regions of
Africa and to
various destinations overseas.  Sasol’s largest export
markets are West Africa and China.  Appellant’s
export
sales are made on a costs and freight (“CFR”), or costs,
insurance and freight basis…
According
to the Commission appellant charges domestic customers of PP
different prices for different grades, and offers customer
various
discounts, but derives all of its domestic prices for PP from an
import parity price (“IPP”) calculation it
performs on a
weekly basis.  In other words, its prices to South African
consumers of PP are IPP-based.  The Commission
contends that
these prices are up to 32% higher than Sasol’s export
prices.’
[6]
According to the Commission, the economic
value of PP sold to domestic customers in South Africa was the price
at which it concluded
its export sales.  This price is excessive
because it bore no reasonable relation to the economic value of PP,
being 32 %
higher than export prices to China and 20% higher than
Hong Kong CFR prices.  The Commission contends that the
difference
was unreasonable and detrimental to downstream consumers.
[7]
In addition, the Commission submitted that
appellant charged excessive prices to Safripol for purified propylene
during the period
2004 – 2007.  Again it was contended
that appellant was dominant in the market for the supply of propylene
to domestic
customers in South Africa, having a market share over the
complaint period of more than 90% of the market.  The Commission

contends that the economic value of propylene can be derived by
taking appellant’s PP export prices, that is the price that
the
Commission considered to be the competitive PP price, and then
deducting a standard margin.   According to the Commission,

this calculation showed that the price for propylene charged to
Safripol was 55% above the economic value.
[8]
The economic value of propylene could also
be measured by predictions made by appellant in 2003 for the price of
propylene.
Appellant had predicted the price of purified
propylene in determining the profitability of opening a new PP plant
(which opened
in 2008) and which was part of the so-called project
Turbo.  Appellant’s actual purified propylene prices
turned out
to be 326% higher than the calculation used in determining
the profitability of Project Turbo.  Accordingly, appellant’s

prices bore no reasonable relation to the economic value and were
detrimental to downstream consumers.    I turn
to deal
in more detail with the relevant products.
Propylene
[9]
Feedstock propylene is a by-product of
liquid fuels produced in an oil refinery.   It is a raw and
impure propylene gas
which has different potential alternative uses,
two of which are relevant to the present dispute.  The first is
to return
the feedstock propylene to the fuel pool where it is then
used in the production of fuel.  The second is to purify the
propylene
for use in the production of PP.   The latter is
a thermoplastic polymer which is a raw material used in the
production
of a wide range of plastic products.  Appellant buys
the propylene feedstock that it uses to produce purified propylene
and
PP from a fellow subsidiary, Sasol Synfuels (Pty) Limited
(Synfuels) at a low price.  Feedstock propylene is produced in
many
oil refineries that employ fuel upgrading processes and by
Synfuels in its synthetic fuels production.  In both cases, it
is a by-product of petrol or fuels production.  Its value is
calculated from its value in alternative uses which involves the

costs of the conversion and the price of that product.
[10]
Synfuels produces liquid fuels from coal.
Feedstock propylene is a bi-product of this process.  As a
result of
a unique process for the production of liquid fuels from
coal, Synfuels suffers a disadvantage when compared to conventional
oil
refineries in the use of its feedstock propylene.
Conventional oil refineries can employ feedstock propylene more
productively
and profitably than Synfuels in the production of fuel,
given their different processes to achieve this end.
Conventional
oil refineries use an alkylation process whereas
Synfuels uses a Cat-Poly process.   The alkylation process
makes more
productive and profitable use of the feedstock propylene
than the Cat-Poly process, with the result that feedstock propylene
is
less valuable in the production of fuels for Synfuels than it is
for the conventional oil refineries.
[11]
Mr McDougall, chief business analyst at the
Sasol Group described the position as follows:
‘
The
important difference between Sasol Synfuels and the conventional
refinery is the value of the feedstock propylene thus derived.

This emerges from the difference in the petrochemical processes.
This difference to a conventional refinery is due to the
absence of
isobutane in the synthetic fuel process, and the substantial volumes
of propylene.
In
the case of Sasol Synfuels, the propylene feedstock stream does not
become part of the petrol pool as an input into high-value
component
such as alkylate.  Instead, it becomes a part of the petrol pool
as a result of being a core input into a Cat-Poly
process.  As
noted above, this plant configuration was the outcome of economic
choices taken at the time of the design of
the synthetic fuels plant
in the 1970’s, in light of the petrol specifications applicable
in South Africa at the time.
The
value of using the propylene feedstock in the fuel-making process is
determined by the net revenues obtained by doing so: this
is the
sales value of the petrol (at the Sasol Synfuels ‘gate’)
less the not inconsiderable costs of the various stages
set out in
paragraph 5.9 above.  Since the factory gate price of petrol and
diesel are very similar on a volumetric basis
and since the costs of
processing the propylene into transport fuels and of hydrogenating
the Cat-Poly product to reduce olefin
levels are essentially the same
for petrol as for diesel, the focus for present purposes will be on
petrol.  Since LPG is
not derived from propylene, its value does
not influence the costs associated with converting propylene to
transport fuels.
This
means that the floor price is the propylene in the propylene
feedstock in the Sasol Synfuels’ process is the price of
petrol
less the significant fuel-making costs (net of the costs of
‘cleaning’ the feedstock propylene stream). This
is the
heart of SP’s ‘special advantage’: Sasol Synfuels
has a fuel production process in terms of which the
opportunity value
to it of the propylene sold to SP is significantly less than petrol,
compared to the conventional refinery in
which the opportunity value
of feedstock propylene produced is significantly more than petrol.
Put
simply, SP is able to purchase propylene feedstock (in chemical
sales) at a value which is significantly less than that which
would
be available from a conventional refinery.  Also, because the
parties are related, Sasol Synfuels does not seek a premium
above
FAV, which an independent party could and would do.  In
practical terms, SP’s Monomers business pays Sasol Synfuels
an
amount for propylene each month, which reflects the fuel prices
during the month in question less a contractually defined value
for
fuel-making costs, since these are the two elements of the FAV
calculation set out above.  Unlike the price of petrol,
the fuel
making element of the calculation does not vary on a month by month
basis, since the elements are set by the initial contract
and the
escalation clauses defined.’
[12]
In summary, as appellant pointed out,
because feedstock propylene is a bi-product of the production of fuel
it does not have its
own costs of production.  It is costed at
an opportunity cost that equals to a value its owner can derive from
it.  The
opportunity cost of feedstock propylene in the hands of
an oil refinery is equal to its fuel alternative value (‘FAV’);

that is the value the refinery can derive from the feedstock by using
it in the production of fuel.   Because of the
relative
disadvantage of Synfuels production process, the FAV of its feedstock
propylene is significantly lower than that of conventional
oil
refineries.   Feedstock propylene in the hands of Synfuels
is worth less than in the case of a conventional oil refinery
because
it cannot use its feedstock propylene as productively and profitably
as can the conventional oil refinery.
[13]
To minimise this disadvantage and to render
the use of feedstock propylene more profitable, Sasol established a
plant for the purification
of feedstock propylene and for the
production of PP.  These facilities were housed in appellant
during the relevant period.
Synfuels supplied feedstock
propylene to SCI and the latter produced purified propylene in its so
called Monomers Division
which sold it to it both to its Polymers
Division and Safripol.   Both appellant, Polymers and
Safripol used the purified
propylene to manufacture PP which was then
sold to buyers both domestically and overseas.
[14]
In short, Synfuels sold its feedstock
propylene to appellant at its own low FAV.   It was
calculated to be extremely low
for the reasons already set out above.
[15]
The Commission contended, notwithstanding
its low costs and the abundance of supply, over demand, that
appellant’s domestic
prices for polypropylene were relatively
high by international standards.   The returns made were
described by the Commission
as “extraordinary”.
According to the Commission appellant’s nett domestic price of
propylene in the
complaint period was at least 12 – 18 % higher
than the domestic prices in Europe.   Over the complaint
period
it enjoyed an average return on capital of approximately 162%
per year and accordingly appellant recovered the full capital
invested
in the business “8 times over”.
[16]
Over the complaint period the Commission
contended that, on the basis of an integrated price, the propylene
and PP businesses considered
together, generated an average return of
62.8% per year.   Thus appellant recovered the full capital
investment in these
businesses 3 times over.
[17]
Accordingly, the Commission contended that,
when appellant’s feedstock cost advantage is compared to the
Western Europe refinery
grade propylene costs, the differences in
domestic prices of PP (excluding appellant’s special rebates)
between appellant
and Western Europe producers was between 41 to 47%
over the complaint period.  In the Commission’s view, the
reason
for this difference was simple: appellant did not pass on any
of its costs advantage to customers, notwithstanding that its supply

exceeds demand in South Africa.   The Commission contended
that the appellant prices its PP with reference to the costs
of
importing PP, which is the customer’s next best alternative.
In its view, these are the highest prices a monopolist
could charge
in a domestic market and substantially higher than the prices at
which appellant sells the same PP which it exports.
In short,
the domestic price of PP was on average 23% higher than its deep sea
export prices.
[18]
This summary of the dispute provides the
background for an analysis of the Tribunal’s decision.
The
Tribunal’s decision
[19]
The decision of the Tribunal,
unfortunately, has proved extremely difficult to understand.
Accordingly, I shall confine myself
to the core findings and eschew
an attempt to parse its reasoning.   After examining the
history of State support for
Sasol, the Tribunal concluded that
appellant’s special feedstock costs advantage had to be taken
into account at some stage
in the enquiry as to the application of s
8 (a) of the Act.  In its view, ‘there is no justification
for the elimination
of the low costs of feedstock propylene from the
evaluation’.  Rejecting the evidence of Dr Jorge Padilla
on behalf
of appellant, the Commission contended that it was not
legally justified to exclude appellant’s special costs
advantage from
the enquiry.
[20]
This conclusion led the Tribunal to an
analysis, in terms of s 8 (a) of the Act, of establishing the price
under scrutiny, comparing
it to the actual costs of appellant and
then deciding if these costs reflected economic value.
[21]
Correctly, it noted that the major
differences between the approach taken by the Commission and
appellant’s experts related
to:
1.
The treatment of feedstock propylene costs,
2.
the valuation of SCI’s capital assets;
3.
the level of capital reward/return on capital;
4.
the allocation of group costs; and
5.
the allocation of fixed costs between domestic and export sales.
[22]
The Tribunal observed that the difference
in the experts’ approaches to feedstock was clearly the most
significant dispute
between the parties.   Central to the
resolution of the case was how the costs of feedstock propylene ought
to be valued.
It was agreed that the feedstock propylene price
should be scrutinised to determine whether the price to be applied in
the analysis
reflected an arm’s length price, being a normal
price under conditions of competition, given that the price of
feedstock
propylene between Synfuels and appellant was between
related firms.  The Tribunal found as follows:
‘
The
appropriate test in s 8(a) analysis is the price charged relative to
the economic value of the good or service in question.
As
highlighted in
Mittal
,
the actual costs of the dominant firm are important evidence of
economic value provided that they reflect normal costs in long
run
competitive conditions.  It is thus necessary to consider
whether SCI’s actual costs properly reflect economic value,

i.e. whether the price actually paid by SCI for feedstock propylene
reflects the price that would have been paid under competitive
market
conditions.
To
determine the above we have specifically had regard to and placed
weight on what Sasol’s own internal and public documents
say
about the price charged by Synfuels to SCI for feedstock propylene.
The
evidence has shown that over the complaint period and most of the
relevant cycle the price for feedstock propylene was determine
under
formulae in supply agreements that were expressly stated to reflect
the FAV of the feedstock propylene to Synfuels.
This pricing
principle is the principle that would apply to sales to all customers
under competitive conditions because it is costs-reflective.’
[23]
By way of an examination of Sasol own
internal documents and communications, the Tribunal thus concluded
that there was substantial
evidence that Synfuels had recognised that
the FAV was the market price of feedstock propylene over a sustained
period.
[24]
Turning to the measurement of appellant’s
capital asset base, the Tribunal noted that the valuation of
appellant’s capital
assets were relevant to the determination
of two items of cost; that is the costs of depreciation, being the
annual costs of the
capital assets used in the business determined by
spreading the total costs of the assets over the use for life and the
return
to investors, being the return to investors for their
investment in the relevant firm.
[25]
The Commission had argued that the
appropriate asset base was the depreciated historical assets values
of appellant’s capital
assets.   By contrast
appellant’s expert, Mr Harman made an adjustment from
historical costs to replacement costs
to account for the impact of
inflation over time and contended that this was the preferable
approach.  An adjustment from
historical costs to replacement
cost reduced the mark ups for purified propylene by approximately
2.8% (tier 1 prices) and 3.2%
(tier 2 prices) and reduced the mark up
for PP by approximately 8-11 %.   The Tribunal accepted
that the replacement
costs was an appropriate proxy for economic
costs in an inflationary environment.   It therefore
employed Mr Harman’s
calculation, save that it rejected
arguments for an additional reduction in the price- cost mark ups by
using insurance values
as an appropriate and reliable value of
replacement costs.
[26]
There was also a considerable dispute
concerning the appropriate return on capital.  Here the Tribunal
was confronted with
two questions: the appropriate risk premium above
the risk free rate and whether this becnhmark is a pre-tax or after
tax measure.
[27]
It appears from the Tribunal’s
determination that it accepted Mr Harman’s calculation, that
the determination of a normal
return on capital must be based on the
computation of the weighted average cost of the capital (WACC) of
appellant rather than
Professor Wainer who contended, on behalf of
the Commission, that the normal return on equity is a risk free rate
plus an equity
risk premium of roughly 4-7% or Dr Roberts’s
equity premium of 3-5%.  However, it rejected Mr Harman’s
proposed
use of an inception WACC and the further addition of a
hurdle rate.  It accepted Mr Harman’s approach that after
tax
return was the appropriate calculation.
[28]
Mr Harman’s proposed adjustment to
group costs to Dr Roberts’ price costs mark ups had the effect
of reducing the mark
ups of propylene by 5.1% (tier 1) and 5.6% (tier
2).  In particular, Mr Harman canvassed three cost items, being
corporate
development costs, Sasol’s management fee costs and
insurance costs.  By contrast, the Commission contended that the

corporate development costs were not necessarily incurred for the
benefit of the purified propylene business.   After

analysing the evidence, the Tribunal came to the view that it was not
necessary to make any adjustments to the price cost mark
ups for
group costs.
[29]
Finally, the Tribunal turned to the
question of the allocation of fixed costs between domestic and export
sales.   The
Commission’s experts had allocated
common costs of the domestic and export businesses in proportion to
the volumes of each
business, there assuming that the fixed costs
were incurred equally across these products.   This
approach was accepted
by the Tribunal which held that the volume
based approach was the most appropriate.
[30]
On this basis, the Tribunal found that the
mark ups of purified propylene prices over actual costs during the
complaint period were
in the range of between 39.9% to 41.5% for tier
2 sales to Safripol and in the range of 25.1% and 26.5% for tier 1
sales for Safripol.
[31]
The Tribunal then turned to other measures
used in the assessment of the economic value of purified propylene;
in particular, the
prices charged by the dominant firm for the same
product in other markets, including export markets and prices charged
by other
firms in other geographical markets.   The
Tribunal adopted the view that an enquiry based upon export prices
for purified
propylene was less reliable than the price-costs test in
order to determine the economic value for purified propylene.
To
the argument that recourse could be had to purified propylene
prices charged by other firms in other geographical markets, the
Tribunal concluded that there were no good comparator in other
geographical markets in relation to purified propylene for the period

under review.
[32]
Turning to polypropylene during the
complaint period, the Tribunal concluded that appellant’s
domestic prices were 41% to
47% higher for respectively homopolymer
and raffia grade compared to the Western European discounted prices
computed on the basis
of feedstock costs comparable to appellant.
Comparing the average export netback price for deep sea exports to
appellant’s
local prices, the Tribunal concluded that the local
prices for PP over the relevant cycle were, on average, 23% higher
than average
deep sea export prices.
[33]
Applying these findings to the wording of s
8(a), the Tribunal rejected the evidence of Dr Padilla that long run
competitive equilibrium
meant conditions of free entry and free exit
from notional competitors in the relevant market in which the firm
under scrutiny
is dominant.  In the Tribunal’s view this
assumption was fundamentally flawed because; ‘
it
disregards the fact that the very basis of the complaint is that in
that market there is never going to be more than one firm
and
therefore never going to be effective competition.  Markets in
which excessive pricing is likely to occur are precisely
those
markets where there never will be entry and exit by new entrants.
’
[34]
In summary, the Tribunal rejected the
argument that appellant should be treated as if it had never received
State support and therefore
that its ‘special advantage”
should not be taken into account.   Accordingly, it held
that, when regard was
had to the price-cost test for PP, there was no
reasonable relationship between the price charged by appellant to the
local plastic
convertors for PP during the complaint period and the
economic value of the product.   Similarly, after examining
the
price-cost test results for purified propylene, the nature of the
product, its importance as an intermediate input in industrial

development, the market characteristics and other circumstances read
within the prism of the objects of the Act, both the tier
1 and 2
prices charged to Safripol during the period bore no reasonable
relation to the economic value of purified propylene.
[35]
It followed that, if a lowered purified
propylene price was charged to Safripol, the latter would be able to
lower its PP price
and offer cheaper PP to the plastic convertors,
expand its PP productive capacity, offer more products to convertors
and provide
more technical service and product development support
and assistance to customers.
[36]
In the view of the Tribunal, consumer
detriment had been demonstrated in relation to appellant’s
pricing for purified propylene.
Similarly, the Tribunal cited
evidence from consumers of plastic products to the effect that the
high price of PP acted to their
disadvantage, particularly in that
local convertors must compete with imports of finished plastic
products.
[37]
Given the nature of the contravention, the
Tribunal ordered appellant to pay administrative penalties of R 205.2
m and R 328.8 m
for the two contraventions of s 8 (a) of the Act;
that is in relation to purified propylene and PP respectively which
were sold
during the complaint period.
Appellant’s
case
[38]
Mr Trengove who appeared together with Mr
Wilson, Mr Gotz, Mr Burger and Mr Marriott, on behalf of appellant,
sought to attack the
foundation of the Tribunal’s reasoning and
thus the Commission’s case that the propylene and PP prices
were excessive
because appellant had failed to pass its unique
feedstock cost advantage on to its customers. Mr Trengove’s
core argument
was that the Tribunal’s reasoning was
incompatible with this Court’s judgment in
Mittal,
supra
.  His argument ran thus:
In
Mittal,
this court sought to engage with the question as to how it can be
objectively determined whether a particular price ‘
has
no reasonable relation to the economic value of the products
supplied’
and may hence be regarded
as excessive.   In doing so, the Court examined the
decision in
United Brands Company v
Commission
[1978] EUECJ C-27/76
;
[1978] ECR 207
and,
particularly, the finding of the European Court of Justice that a
mere comparison of prices at which the seller actually sold
the
relevant product to different buyers in the same market was an
insufficient basis to conclude that the higher price was “excessive”,

even where the price was 50% higher than the lower price. (paras 261
- 268 of
United Brands
)
[39]
Pursuant to this, the Court in
Mittal
said the following at para 40:
‘
What
the legislature must be taken to have intended by ‘economic
value’ is the notional price of the good or service
under
assumed conditions of long-run competitive equilibrium.  This
requires the assumption that, in the long run, firms could
enter the
industry in the event of a higher than normal rate of return on
capital, or could leave the industry to avoid a lower
than normal
rate of return.  It does not imply perfect competition in the
short-run, but rather competition that would be
effective enough in
the long run to eliminate what economies refer to as ‘pure
profit’ – that is a reward of
any factor of production in
excess of the long-run competitive norm which is relevant to that
industry of production.’
[40]
Mr Trengove noted that, at para 43 of the
Mittal
judgment, the Court held that economic value ‘is a notional
objective competitive, market standard, and not one derived from

circumstances peculiar to the particular firm.  Thus where the
firm’s prices are no higher than economic value, no
contravention of s 8(a) can arise.
[41]
If the firm’s prices are higher than
economic value so determined, the test of reasonableness in respect
of the difference
remains to be applied.  Significantly for Mr
Trengove’s argument  the court held as follows at para 43:
‘
The
test of reasonableness applies to the excess of price over economic
value, and thus only to the element of ‘pure profit’

(over and above ‘normal profit’) implicit in that price.
It is at this stage of the enquiry that circumstances
peculiar to the
particular dominant firm would rationally come into the reckoning.
It would seem sound, when considering
whether the higher price bears
a reasonable relation to economic value or not, to take into account
the benefits flowing to the
firm the subsidised loan, long-term low
rental, or other special advantage which may serve to reduce its own
long-run average costs
below the notional norm.’
[42]
In seeking to tease out the implications of
the Mittal case, Mr Trengove referred to the evidence of appellant’s
expert, Dr
Padilla.   In his testimony, Dr Padilla referred
to paragraph 40 of
Mittal
and noted that this passage did not equate economic value to costs
and certainly not to the costs of a dominant firm.  For
Dr
Padilla, the challenge was to determine normal profit in terms of a
level that would be consistent with the long run competitive
norm.
Dr Padilla noted that competition eliminates rents that are not firm
specific:
‘
Firm
specific cost advantages are not transferred to consumers by the
competitive process.  Firm specific rents associated
to firm
specific advantages are compatible with the long run competitive
norm.  The CAC’s definition of economic value
makes
perfect sense as a matter of economics …
.
And it makes sense
because free entry … because it talks about free
entry and
free exit.  And by talking about free entry, the court is
telling us this price that we regard as the economic value
is a price
that serves the interest of consumers.  Interest of consumers
are protected, because if the dominant company tries
to raise prices,
entry will discipline that firm, it will bring prices to a level that
is optimal for consumers.
But
at the same time the court is imposing the free exist condition and
he’s telling us, but we are not crazy.  We don’t

want to go to the point of extracting all costs advantages so that
firms are losing money.  The free exit condition tells
us we can
transfer rents to consumers, but there is a limit.  And the
limit is that it will go beyond a certain level then
these firms will
exit the market.’
[43]
Turning to the question of how long term
competitive equilibrium is to be calculated, Dr Padilla contrasted
the approach of the
Commission to that of appellant:
‘
The
Commission’s approach relies on two assumptions.  The
first assumption is that the entrants are clones of Sasol.
The
free entry process and free exit process in the Commission’s
case is what I term later a world of clones.  The second

assumption is that entry and exit doesn’t happen in South
Africa. It happens in a market that is many, many, many times larger

than South Africa.  I will explain later that it has to be a
market of galactic proportions.  And then, you know, it
makes
sense, the clone wars in a galactic market.
My
approach is completely different, I’m focussing on entrants
that are efficient, but that they’re efficient, having
access
to the technology that is available in the market, not the technology
that Sasol has, which is unique, which is peculiar,
but this time
peculiar, not in the sense of unusual, but in the sense of the Oxford
dictionary that it belongs to and comes from
the Latin term,
peculiars, private property.  It is the private property, that’s
why it is peculiar.  My entrants
don’t have that peculiar
technology.  My entrants have the technology that is available
in the market, their standard
refineries.’
[44]
On the basis of this evidence, Mr
Trengove submitted that the notional competitive market price is
based on the costs of the notional
competitor.  Accordingly, in
an excessive pricing case, the actual costs of a dominant firm may
constitute useful evidence
of the costs of the notional competitor,
but the actual costs of a dominant firm are not the benchmark. The
benchmark remains the
costs of the notional competitor.  The
dominant firm’s actual costs may only be used if, and to the
extent, that these
costs reflect those of a notional competitor.
On this basis, Mr Trengove argued that appellant’s peculiarly
low
feedstock costs had to be disregarded in the determination of
economic value of propylene and PP which, ultimately, stood to be

based on the costs of feedstock propylene in a competitive market.
[45]
Mr Trengove contrasted this approach to the
evidence of Dr Roberts, on behalf of the Commission, particularly the
latter’s
approach to the decision in Mittal in general and to
paragraph 43 thereof in particular.  Under cross examination, he
put
it to Dr Roberts that a long term competitive norm is the
benchmark by which costs are to be determined.  Costs that did
not
conform to this standard, that is costs which are special to a
particular firm are to be disregarded.  By contrast, Dr Roberts

suggested that, in a hypothetical enquiry test for the long term
competitive result, notional competitors had to be postulated
as
having the same cost structure as the dominant firm.  The
following exchange is illustrative:
‘
ADV
TRENGOVE:   You say that the notional competitor, one
should postulate, should have the same costs as the dominant
firm?
DR
ROBERTS:          Apart
from if they are peculiar.
ADV
TRENGOVE: Well what renders a costs peculiar if you simply replicate
it and attribute it to every competitor in the notional
market?
DR
ROBERTS:          Well I
think there are a number of things that one, which are
not
necessarily affected in the decision, but whether it’s due to
risk taking or innovation.  This what patents protect.
So
if you are engaged in peculiar things that were particular to the
firm in terms of risks that is made, then those will clearly
be
things that one would take into account as being peculiar to that
firm, which would be different from just being located close
to or
having the abundant feedstock, I mean locate in a place which allows
access to that feedstock.
ADV
TRENGOVE:     So you understand the costs that
ought to be disregarded, not merely as all of the costs that
are
peculiar to the firm, you only exclude certain categories of
peculiarity?
DR
ROBERTS:          It’s
a definition of peculiarity, which distinguishes between,
I believe,
the costs of SCI in terms of the feedstock costs, abundant low costs
feedstock and other things that could be peculiar.’
[46]
Mr Trengove placed emphasis on Dr Roberts’
acknowledgement that economic value must be determined not only on
the basis of
firm specific costs but also by taking account of the
firm specific history of the firm under investigation.
According to
Mr Trengove, Dr Roberts pursued a conclusion that would
yield a specific economic value contrary to the holding in
Mittal
:
‘
DR
ROBERTS:          I said
things that … sorry, maybe we can dominate to say, the
things
that would be regarded as peculiar would be things such as particular
innovation, risk taking, behaviour.  Things which
are …
and again this potentially a disagreement with the CAC, because they
talk about subsidised loans or lower market rentals,
so I find it
difficult to understand the context in where that comes from.
But I’m suggesting that if there particular
innovation or risk
taking by the firm, then this would be something that they would have
a return on.
ADV
TRENGOVE:     But those … the court’s
language there was quite clear.  It was language
sued in the
context of Arcelor Mittal, which had a history of state support.
DR
ROBERTS:          Well
to be very blunt, this is why I find it unusual, because
subsidised
loans didn’t come into that case.   What came into
that case was lower costs iron ore and they make
no mention of that
here.   So I agree with you, it’s … you would
expect it to be in relation to the case,
but it actually doesn’t
mention the main thing that was peculiar in the sense or special to
Arcelor Mittal, which was the
25 years of costs based iron ore.
ADV
TRENGOVE:    It spoke of cost saving to the firm
resulting from a subsidised loan or lower than market
rental, or
indeed any other special advantage.  That doesn’t mean
only advantages, deserving advantages as you suggest,
it means any
advantage, current or historical?
DR
ROBERTS:          On
that interpretation, then I disagree with the CAC decision.
So
if you were to ask me I agree with all of it, then on that
interpretation I definitely disagree with the CAC decision.’
[47]
In Mr Trengove’s view by following
the approach of Dr Roberts, the Tribunal had, in effect, repudiated
the fundamentals of
the
Mittal
holding.  In this connection he referred to paragraph 95 of the
Tribunal’s determination:
‘
The
Court (in
Mittal
)
did not disregard this specific example of costs advantages.
The context however was a particular approach of determining
in the
first instance on a notional level what competitor’s costs
would be in a notional competitive market.  The Court
was
concerned with a far broader and holistic approach.
Furthermore, we must have regard to all relevant factors because

ultimately one is trying to determine whether in a particular case
the price charged in a particular environment and in particular

circumstances was excessive.
’
From this, the
Tribunal came to the following key conclusion:
‘
Where
the dominant firm’s position in a particular market is not the
result of any innovation or risk-taking on its part but
rather due to
current or past exclusive or special rights, one therefore would want
to have regard to those facts.  Thus,
part of the s 8(a) enquiry
should be an explanation for why the dominant firm is able to charge
a price above the economic value
of the good or service in question –
in particular, if this ability is the result of its own efforts (for
example, risk taking
or innovation), so that the high prices should
be regarded as an appropriate reward for the firm’s competitive
efforts,
or if it is simply the result of the firm taking
advantage of its entrenched dominance, in which case its actions, to
the extent
that they harm consumers/customers, may be an abuse as
contemplated in s 8(a).’
[48]
In Mr Trengove’s view, this finding
meant that economic value is no longer to be regarded an objective
market standard applied
across the board to all dominant firms.
It is a subjective standard in the sense that it is based on a
subjective value
judgment of the specific firm in the sense that it
is predicated on the peculiar history and costs of a single firm and
applies
only thereto.
[49]
It was on this basis that the Tribunal
rejected Dr Padilla’s approach in that it held if ‘one
excludes SCI’s special
costs advantage from the first stage of
enquiry, regardless of its origin, one will never take that advantage
into account’.
In other words, on Dr Padilla’s
interpretation one must engage in a notional exercise and, if the
result of that produces
no difference between the notional economic
value and the actual price, then the enquiry stops.   This
leads to an artificial
result and is a misreading of the
Mittal
judgment when applied in its broader
context.
[50]
Mr Trengove vigorously attacked this
finding, contending that it represented a flagrant violation of the
doctrine of precedent,
in that it was so obviously a misreading of
the
Mittal
judgment.   In his view, the Court’s interpretation
of s 8(a) in
Mittal
gives effect to its purpose; that is it made the benchmark against
which the dominant firm’s price stands to be judged the

notional price that would have prevailed in a competitive market.
It posits competition and then asks how the dominant firm’s

price compares with the price that would have prevailed in a
competitive market.  It compensates for the lack of competition

in the market by assessing the dominant firm’s price against
the price that would have prevailed in a competitive market.

In Mr Trengove’s view, the purpose of s 8(a) and therefore the
prohibition of excessive pricing does not seek to achieve
anything
more than to compensate for the lack of competition in the market.
On the basis of what he referred to as
the vague subjective and
firm specific rule adopted by the Tribunal, it would be impossible
for any dominant firm to determine
where a permissible price ends and
an impermissible price begins.
[51]
The importance of the
debate about feedstock was highlighted by Mr Trengove’s
submission that, even if the Court accepted
all the Tribunal’s
findings and figures, save to adjust appellant’s feedstock cost
to a competitive market value, the
premium of appellant’s
prices over economic value declined to 7.2% for propylene and 0.8%
for PP.   There would
then be little left of the case
against appellant as these prices would manifestly be reasonable in
relation to economic value.
Respondent’s
case
[52]
Mr Wesley, who appeared with Ms Lekoane on
behalf of the Commission, (the heads on behalf of respondent having
been prepared by
Mr Subel, Mr Wesley and Mr Lekoane) contended that
the only sensible approach to the determination and the consequence
of a long
term competitive equilibrium was to consider the notional
competitive market as one where pure profit has been competed away
and
where price reflects costs.  If the notional competitive
standard was considered to be one where pure profit was competed
away, what was left to be considered was a market, where there
existed competition between efficient firms which were identical
to
the dominant firm.  Given the scale of the economy in propylene
and PP, this would imply that a long run competitive equilibrium
had
to be considered in a somewhat larger market that the South African
market.
[53]
Mr Wesley also submitted that to consider
outcomes under a notional long run competitive equilibrium in order
to derive economic
value requires an assumption that the rival firms
have the same or lower costs of feedstock as appellant because
Synfuels had abundant
feedstock propylene, and, absent exerting its
monopoly power, should supply the notional competitors at its FAV.
The
feedstock is not “special” to appellant, apart from
appellant not being subjected to the full exertion of monopoly power

by Synfuels which should not fall within the meaning of a “a
special cost advantage”.
[54]
Mr Wesley submitted that the only cost
advantages which should be taken into account in favour of a dominant
firm in an excessive
pricing enquiry are those that are the product
of risk and innovation by the firm.  It is only where there has
been such risk
or innovation that a firm might be entitled to charge
high prices as a consequent reward.  But a large difference
between
the price of actual costs might still not be reasonable.
In his view, the effect of appellant’s case in general
and the
evidence of Dr Padilla in particular was that there would be a
profound distortion of the approach to an abuse by way of
excessive
pricing.    Dr Padilla assumed that any firm’s
lower costs was a product of efficiency of the firm.
He did not
consider that a cost advantage might not be the product of a firm’s
own efforts but simply the result of previous
state largesse,
precisely the situation when an abuse by way of excessive pricing
might well occur.
The
differences regarding the dispute with regard to the interpretation
of
Mittal
[55]
To return to the facts: It appears that
Synfuels sold propylene feedstock to appellant under three sets of
agreements.   In
1994 there was an agreement for a stream
of propylene known as the condensate 3 stream.   In 1999
another agreement was
concluded for additional feedstock known as the
condensate 2 stream and in 2003 a 1 tier agreement was reached which
governed the
supply of all feedstock by Synfuels to appellant.
[56]
The actual price that appellant paid during
the entire complaint period was the price set out in terms of this
latter agreement.
As noted earlier, the Tribunal used
appellant’s actual costs of feedstock propylene, that is the
price in terms of the 2003
agreement in its price – cost tests
for both propylene and PP.  It did so on the basis that ‘
under
conditions of competition, Synfuels would sell its feedstock to all
customers in the South African market at its uniquely
low FAV because
it is cost reflective.’
[57]
Mr Trengove submitted that a firm with a
unique cost advantage would have no incentive to reduce its price
down to its own costs
as it could effectively compete at its
competitor’s costs or capture the entire market at a price only
marginally lower.
The appellant’s unique costs advantage
meant that its costs were lower than those of its competitors and it
would not have
to compete down to its own cost.  It could reduce
its price just below its competitor costs but no further.
Accordingly,
the Tribunal’s findings that ‘under
conditions of competition FAV is the price which Synfuels would sell
its feedstock
to all customers in the South African market because it
is cost - reflective’, would hold true only in circumstances
where
all firms are able to supply feedstock propylene to appellant
at Synfuels’ FAV.  This could only happen if the upstream

market for the supply of feedstock propylene was populated by a
multitude of clones of Synfuels and would all share its uniquely
low
FAV.  By contrast, appellant argued that in a hypothetical
market populated by Synfuels and notional typical suppliers
of
feedstock (that is conventional refineries) the price would tend
towards the marginal refinery’s alkylation FAV.
By
contrast, Synfuels was the only supplier of propylene feedstock, had
no competition, and buyers would thus compete the price
up to their
breakeven level.
[58]
For all these reasons, Mr Trengove
submitted that the Tribunal had failed to appreciate that the FAV
concerned was the alkylation
value of the feedstock.  The
opportunity costs of ordinary refineries served as a floor price for
feedstock propylene.
But, if Synfuels was to behave
rationally, it would only need to price at a level just below the
opportunity costs of the ordinary
producer to capture the entire
market.  In turn, this meant that any conclusion that
appellant’s actual feedstock costs
reflected a competitive
market price was wrong.
[59]
Mr Trengove also referred to documentary
evidence which had been made available to the Tribunal, in particular
the three agreements
to which I have made reference.  They
reflected that Synfuels had, from time to time, negotiated for the
sale of feedstock
to third parties.  In all of these
negotiations, it had consistently demanded more than its FAV and was
never prepared to
sell to third parties at a price equal to its own
FAV.  Thus, in the 1994 agreement, which was part of a basket of
agreements
between Sasol and AECI which established the Polyfin JV,
Synfuels undertook to supply feedstock to Polyfin for which Sasol
would
receive FAV together with a share of the joint venture.
In Project Mango, Synfuels demanded its 1994 feedstock price plus
$45
to correct for errors in the 1994 agreement and a premium of $60.
In Project 2003, Synfuels offered to supply feedstock
to the JV but
at a premium above its FAV of 50% of the difference between its price
and the next lowest price.
[60]
According to Mr Trengove, by rejecting this
evidence, the Tribunal had failed to appreciate that the significance
of this evidence
lay in the fact that it supported the economic
theory upon which appellant had based its case.   Synfuels
might surrender
its unique feedstock costs advantage to its fellow
subsidiary, being appellant, but it did not have an incentive to do
so in relation
to third parties.  This conclusion reflected how
a producer with a unique feedstock costs advantage would act in a
competitive
market.
[61]
In summary, appellant’s case was that
its actual feedstock costs were not those which have been incurred by
a notional producer
in a competitive market. It enjoyed a special
cost advantage which meant that its costs had to be adjusted upwards
in accordance
with the principles of the decision in
Mittal
.
Mr Trengove noted that the Commission, which bore the onus of proof,
had never established the nature of this adjustment
and had made no
attempt to do so, save to attack the appellant’s recourse to
the refinery floor price.
[62]
Mr Trengove submitted that the erroneous
treatment of appellant’s feedstock cost advantage not only
inflated the margin above
“economic value” but also
contaminated other features of its decision.  In particular, it
was only by using appellants
very low cost of feedstock propylene in
its calculations of economic value that the Tribunal was able to
obtain its results for
its price-cost test for propylene.  It
was only by artificially decreasing appellant’s price for
propylene as an input
cost that it had obtained the results it did
for PP and its price-cost test.  Further, the Tribunal justified
its dismissal
of the relevance of the propylene price comparators in
other geographical markets solely on the basis that the foreign
producers
did not enjoy appellant’s feedstock costs advantage.
Only by artificially decreasing PP prices in Western Europe
to
account for this cost advantage, could the Tribunal have concluded
that appellant’s PP prices was substantially above
the foreign
prices.
[63]
By contrast, Mr Wesley submitted that the
feedstock prices contained in the supply agreements were actually
higher than Synfuels’
actual FAV.   The prices under
both the 1999 and 2003 agreements did not reflect the FAV because
they did not correctly
measure the price of fuel that otherwise would
have been made from the feedstock nor the costs to make and supply
fuel from feedstock
propylene.
[64]
The Commission disagreed with the
Tribunal’s finding where it had not accepted the Commission’s
downward adjustment
to take account of FAV in its profit cost
analysis.
[65]
In summary, the Commission contended that
the prices under the 2003 agreement were some 9% higher than
Synfuels’ actual FAV
during the complaint period.  For
this reason, the Commission submitted that any analysis of economic
value based on SCI’s
costs should include a downward adjustment
to those costs to reflect actual FAV.
[66]
The Commission contended further that the
price for feedstock propylene in the 1994 agreement, which also
purports to be at Synfuels’s
FAV, is 15% lower than the price
under the 2003 agreement over the period FY02-FY08.  Initially
in his witness statement Mr
MacDougall attributed this difference to
missing depreciation costs.  To make an allowance for this in
its initial calculations
the Commission, added an amount of R 100 per
ton (inflated yearly) to the 1994 prices when performing its analyses
of economic
value based on actual FAV during the complaint period
(the “1994 + R100” estimates).
[67]
During his evidence, Mr MacDougall
indicated that the difference between the 1994 prices and the 2003
prices were that various costs
were included or included in the
different agreements, or measured differently.  In the view of
the Commission it was notable
that, he made no mention of the
depreciation costs that he had relied in his witness statement, as
explaining the difference.
[68]
Using Mr MacDougall’s new evidence,
as well as evidence emerging from documents discovered by appellant,
the Commission claimed
to have performed a direct analysis of whether
the 2003 prices reflected actual FAV, having regard to the different
costs and measurements
of costs used in calculating those prices.
[69]
The Commission contended that this analysis
supported its initial conclusion that the prices charged by Synfuels
under the 2003
agreement did not reflect actual FAV.  In fact,
on the Commission’s revised analyses, actual FAV was below the
“1994
+ R 100” estimate it had initially used.
[70]
This Court does not need to
engage in this particular debate unless it is satisfied that the
price-cost analysis based on the FAV
is in itself not justifiable in
terms of s 8 (a) of the Act.  It is to this question that I must
now turn.
The
proper application of s 8(a)
[71]
As Mr Wesley correctly noted there are two
fundamental debates between the parties in this appeal in relation to
the application
of s 8 (a), read together with  definition of
excessive price as defined in s 1(1) (ix) of the Act.  The first
relates
to the proper interpretation of the phrase ‘economic
value’ and the second to the manner in which the reasonableness

of the relation being price and economic value is to be assessed.
[72]
I deal first with the question of the
resolution of the debate with regard to economic value.
[73]
Under cross examination, it was put to Dr
Roberts that the Commission must succeed in the feedstock debate in
order to win its case.
Although he equivocated (as
unfortunately was the case under much of his cross examination), Dr
Roberts did concede that the basic
question of the feedstock cost was
‘
by far and away the biggest’
dispute between the parties.
[74]
Mr Trengove asked Dr Roberts if he
calculated the costs of propylene, made his comparisons and then
arrived at the conclusion that
the propylene price was excessive in
that it exceeded the economic value at a margin of between 36 to 71
%.   This set
of figures had emerged from the supplementary
economic expert report filed by Dr Roberts in which the following
appeared:
‘
Adjusting
the costs to reflect the FAV based on Sasol Synfuels’ actual
opportunity cost and applying bond +3% return, I find
that prices
charged to Safripol for purified propylene  (Tiers 1 and 2) are
on average 36% and 53% (Sasol calculated FAV,
condensate 3 contract +
R 100), 45% and 64% (discount OOC fuel sales) and 52% and 71%
(coastal sales) higher than total cost of
production, for FY2002 to
FY2008.’
[75]
Dr Roberts conceded that this ‘excess’
depended ‘vitally on the feedstock debate.’  The
following exchange
clarified this issue:
‘
ADV
TRENGOVE:    It is only when one takes account of the
feedstock advantage, in other words, postulates the low
costs as the
cost on which the economic value has to be determined that it jumps
from 10% to 33%.
DR
ROBERTS:          That’s
correct.
ADV
TRENGOVE:     And then the range from 33 to 59 is
a range which depends on the particular alternative value
one
attributes to the feedstock.
DR
ROBERTS:
Agreed.’
With
regard to PP the following exchange is instructive:
‘
ADV
TRENGOVE:    This is an outcome reached by again doing
the calculation of cost from the ground up.  Correct?
You
call it an integrated approach.
DR
ROBERTS:          Yes,
that’s right, ja.
ADV
TRENGOVE:     You don’t assume you don’t
make any assumption about the costs of propylene as
an input cost.
DR
ROBERTS:          No.
ADV
TRENGOVE:     This is again a calculation right
from feedstock cost through to price.
DR
ROBERTS: Yes.
ADV
TRENGOVE:     And substantially it is the same
excess, which you say exist in the propylene price, which
is being
replicated in the polypropylene price.
DR
ROBERTS:          I agree
entirely, ja.
ADV
TRENGOVE:     And again vitally dependent on the
feedstock assumption.
DR
ROBERTS:          Agreed.’
[76]
Dr Roberts then developed his explanation
by pointing out that the excessive pricing assessment which was
conducted by the Commission
was the excessive pricing by a dominant
firm; that is appellant and not Synfuels.  Accordingly, what was
required was to examine
the dominant firm’s costs in terms of
the competitive norm ‘
so, this would
be where the firm is efficient or inefficient, for example, in terms
of its average costs.  So, you got to look
at well are the cost
of the firm, do they correspond to the competitive norm?  We are
talking about the costs of its production
processes and you got an
exercise there that you have to do and that is where we believe he
has an onus to say that SCI is efficient.
’
[77]
By contrast, Dr Padilla based his report on
two readings of the
Mittal
judgment.
In the first instance, his results were based upon that which he
referred to as the Mittal 1 test that the notion
of economic value
was the notional price of a good or service under assumed conditions
of long term competitive equilibrium.
This equilibrium arises
when there is free entry and free exit.  In the present case,
the Mittal 1 test would relate to a
situation where the seller of
feedstock propylene had low opportunity costs and bargaining power
and the buyers of feedstock propylene
did not have any bargaining
power because of the fact that they were in a long term competitive
equilibrium.  According to
Dr Padilla, under the Mittal 1 test,
the relevant market price for feedstock propylene would be one in
which downstream propylene
producers would just breakeven in economic
terms.  It assumed that Synfuels was in a position of bargaining
power because
it was the only supplier of feedstock propylene in
South Africa, possessed of the advantage of the Fischer-Tropsch
process; that
is the process by which oil is produced from coal.
[78]
Central to the debate about a price cost
test, was the so called Mittal 2 reading.  The Mittal 2 test
envisaged that economic
value be assessed using the dominant firm’s
economic costs of production.  But on Dr Padilla’s
reading, the Court
in
Mittal
made it clear that the dominant firm’s own costs will only
provide a measure of economic value if they correspond to the

economic norm.  Accordingly, he argued that adjustments to the
dominant firm’s costs may need to be made to arrive at
an
estimate of its costs, and thus economic value, in the notional
competitive norm.
[79]
In adopting a price-cost
comparison approach, Dr Padilla testified that price-cost margins
should be adjusted for firm specific
costs to avoid penalising
efficiency.
[80]
Central to the debate then between Dr
Roberts and Dr Padilla, as interrogated through the cross
examination, of Dr Roberts was Dr
Padilla’s postulate that the
determination of the cost of feedstock would be equivalent to the FAV
of the conventional South
African refineries.
[81]
In support of its contention that its
actual feed cost were not that which would have been incurred by a
notional producer in a
competitive market and that it had a special
cost advantage, appellant sought to adjust its cost calculation in
accordance with
its reading of the Mittal case by reference to a
quantification of the South African refinery floor price.
[82]
This calculation was provided by Dr Koster
who sought to calculate the ‘so-called refinery floor price’
which was not
designed to allow the refinery any margin over the
propylene – in – alkylate valuation  or additional
costs associated
with operating a depropaniser which the oil refiner
will need to install to separate refinery grade propylene from the
fluidize
catalytic cracking gas stream.   According to Dr
Koster:
‘
The
annual difference between the refinery grade propylene prices and the
propylene-in-alkylate value was calculated for each of
the years
between 1998 and 2007.  The low points from this calculation
were then averaged, with the assumption being that
the low points
represent market clearing conditions, with no margin over costs.
This average was found to be US$45/ton.
This value was taken to
be the minimum mark-up a refinery operator is prepared to accept to
divert propylene away from the alkylation
unit.
’
[83]
Dr Koster compared the transfer price that
appellant paid to Synfuels for the feedstock propylene to the
refinery floor prices calculated
the found that over the period of
July 1997 to December 2008 the propylene – in – alkylate
value exceeded the transfer
price by 24.5%.  On this approach,
appellant’s business enjoyed a price advantage of 24.5% held at
the marginal and
arguably the next ton of propylene that would be
produced in South Africa.
Evaluation
[84]
The key question thus arises: is it
permissible to ignore appellant’s lower costs, which it enjoys
as a result of its relationship
with Synfuels which produces an
excess of propylene as a result of its specific technology and
construct a price which increases
the costs of appellant based on a
postulate of a competitive economy in which the cost of feedstock
would be significantly higher
than is the case where there is of the
alleged special advantage enjoyed by appellant.
[85]
In summary Dr Padilla contends as follows:
Synfuels produces large quantities of propylene feedstock and has low
opportunity costs
measured by the FAV.  Appellant benefits from
the low costs feedstock due to its vertical integration with Synfuels
and Synfuels’
unique technology.  Dr Padilla contends that
Synfuels, like any other commercial enterprise, would not transfer
all of the
benefit of its low costs in an unrelated chemical
business. Hence, the market price of propylene feedstock would not
therefore
be equal to FAV.  This is demonstrated by what is
referred to as Project Mango.
[86]
Briefly the background to this
project was that in 1996 Polifin Limited and Dow, which was then
rated as one of the world’s
leading polymer companies, began
discussions on establishing a new PPJV, a project known as Project
Mango.  The envisaged
co-operation aimed at expanding PP
production at Secunda by a further 200 000 tons per year over
and above 80 000 per
year expansion envisaged in what was
referred to as a PP phase 2 expansion project.  Of relevance to
the present analysis
is that the Mango JV would have required
additional feedstock propylene.  Polifin therefore approached
Sasol for feedstock
supply.   As Sasol was planning to use
its feedstock propylene in an alternative chemicals project
(oxo-alcohol), it
offered a feedstock propylene price to Polifin
which was set so that Sasol would be no worse off with Project Mango
than with the
so-called oxo alcohol project.
[87]
This first offer was at US $ 376 per ton
Polifin convinced Sasol that a contract with Polifin would be less
risky while the oxo-alcohol
project would carry some risk.  This
price was decreased to $ 330 per ton but at this price Project Mango
was considered not
to be viable.
[88]
According to various Polyfin board minutes,
it appears that Synfuels position was that the price would at least
have to be the Condensate
3 price plus US $ 45 per ton, plus US $ 60
per ton opportunity value determined for the proposed oxo–
alcohol project.
[89]
Dr Padilla considered that this constituted
evidence that the market price for propylene feedstock would not be
equal to the FAV
of Synfuels.  Taking the present case where
there was only one seller of propylene feedstock with lower
opportunity costs,
being Synfuels, and that there was a long run
competitive equilibrium assumed at both propylene and PP levels, Dr
Padilla contended
that the correct approach was to adjust actual
costs upwards.  His preferred approach was to adjust the
propylene feedstock
propylene price to the South African refinery
floor price.
[90]
This calculation stood in sharp contrast to
the Commission’s approach that the transfer price should be
adjusted downward
to what it took to be Synfuels’ opportunity
costs for supplying propylene feedstock to appellant.   In
his first
report, Dr Roberts had contended that the feedstock costs
were equivalent to that which was contained in the 1994 feedstock
agreement
which formed part of a basket of agreements by which Sasol
AECI established their Polyfin JV.   In terms of these
agreements
Synfuels undertook to supply feedstock to Polifin under
this feedstock agreement for which Sasol received FAV, together with
a
share of the JV.  This approach was then altered and Dr
Roberts argued that these costs was lower than those obtained from

applying the 1994 agreement because, during the complaint period,
Sasol Oil had discounted all exported fuels at lower prices.
[91]
Dr Padilla attacked this approach on the
basis that it was inconsistent with long run competitive equilibrium.
As he stated:
‘
the only way in
which competition between Synfuels and the hypothetical companies
could bring the price of feedstock to the costs,
the fuel alternative
value for Synfuels, is if you pass a free entry, free exit model for
the clones in the market that is far
bigger than the South African
market.
’
He further contended
that to base the entire costing on the 1994 agreement would be to
ignore the context of the negotiations between
the parties which
culminated in the agreement and, in particular, the allocation of
shares in Polifin between Sasol and AECI. He
further contended that
the Commission was asking this court to accept that the 1994
agreement could be projected forwards to 2004
to 2007 without any
adjustment.
[92]
The comparison between the Commission’s
approach as advocated by Dr Roberts and the approach of the appellant
as set out by
Dr Padilla was that Dr Padilla adopts as the costs of
the feedstock propylene an amount of R 5465 per ton.
Appellant’s
actual costs, that is the price it paid to
Synfuels for the feedstock was in the amount of R 4453 per ton.
By contrast,
the Commission invoked three prices, the FAV based on
the 1994 agreement, of R 4076 per ton, the FAV based on OOC discounts
of
R 3881 per ton and the FAV based on coastal sales of R 3701 per
ton.
[93]
Given that the Tribunal did not accept all
of Dr Roberts’ testimony, it is best to focus the evaluation on
its finding which
was based, it said, ‘under conditions of
competition, FAV is the price at which Synfuels would sell its
feedstock to all
customers in the South African market because it is
cost reflective’.  On this basis, it used appellant’s
actual
costs of feedstock propylene, that is the price set out in
terms of the 2003 1 Tier agreement in its price cost test  which

it applied for both for propylene and indeed for PP.
[94]
There is certainly evidence elicited under
cross examination by Dr Roberts that supports this approach.
The importance of
this evidence necessitates reproduction of the full
exchange:
‘
ADV
TRENGOVE:    Now, the floor of these producers,
firstly, if one were to take the other refineries, not SCI, they

would, their floor, their minimum price of each of them would be its
fuel alternative value.
DR
ROBERTS:          I
agree, ja.
ADV
TRENGOVE:     None of them would reduce their
price below their fuel alternative value.
DR
ROBERTS:          No,
their real fuel alternative, yes.
ADV
TRENGOVE:     Indeed, but whether they compete
down to that level or not depends on whether they …
it depends
on their output and the dynamic in the market.  If they can
dispose of their output without lowering their price
to the floor,
they will do so.
DR
ROBERTS:          Yes
that follows.
ADV
TRENGOVE:     Synfuels will compete in that same
market and it will also lower its price as far as may be
necessary to
dispose of its output, as all the others would.  Correct?
DR
ROBERTS:          If it
is selling as effectively a monopoly, yes it wouldn’t…
ADV
TRENGOVE:     Well, it’s not a monopolist.
It’s a competitor for as long as the …
it has
competition for as long as the floor, for as long as the price is
above the floor of the SA refineries. Correct?
DR
ROBERTS:          It
just says input.  This would be, as long as there is excess

capacity in terms of the other refineries, but if they are capacity
constrained, then actually they are irrelevant.
ADV
TRENGOVE:     But for as long as the price is
above the SA refineries  price, it would compete and
it would be
in competition with the other refineries.
DR
ROBERTS:          No,
no, if the other refineries, and this is where capacity, they
comes
back in, if the other refineries have hot very minimal capacities,
they could choose to sell their refinery floor prices
and Synfuels
could charge much higher prices.  I mean, Synfuels is a
monopolist.
ADV
TRENGOVE:     Well, capacity constraints will
allow anybody to charge any price, if the capacity is sufficiently

constrained, but it allows, assume for the moment, sufficient
capacity.
DR
ROBERTS:          So,
all the other refineries have got excess capacity, then they
would
set the price, ja.  They would be the ceiling to Synfuels market
power.
ADV
TRENGOVE:     Synfuels would also reduce its
price sufficiently to dispose of its output, but there is one

difference and that is…
DR
ROBERTS:          Well,
that’s not true.  I mean, that’s not true.
If
it is exerting market power, it will only reduce its price according
to the demand of the buyers, not dispose of all its output,

absolutely not…
ADV
TRENGOVE:     What I’m saying is it will
reduce its price only insofar as may be necessary to dispose
of all
its output.
DR
ROBERTS:          No, if
it is necessary to dispose of all its output, then it is
different.
If it is saying here are the refineries, they’ve got a price
and we are competing down to that price,
they are not going to keep
on pushing the price down until they’ve disposed of its
output.  Then they would have to
go and provide prices that
allow the buyers to increase their sales to meet…  so
that the buyers, because Synfuels
essentially has got to derive
demand.  It derives demand from the buyers.
If
it is going to dispose of all of its output, then it is going to have
to reduce its price below the other refineries in order
to allow to
its buyers to increase their supply, their production and meet other
markets.   So, it would reduce its price
presumably as a
monopoly, just until it hits the ceiling that’s provided by the
refineries.
ADV
TRENGOVE:     Yes, the point is as soon as that
price reaches what we’ve called the SA refinery floor
price, do
you understand what I mean by that term?
DR
ROBERTS:          Yes.
ADV
TRENGOVE:     As soon as the price, the
descending price hits that level, then below that level SCI is the

only seller
.
DR
ROBERTS:          Ja.
ADV
TRENGOVE:     Which would mean that it would have
no incentive to compete, to lower its price down to its
own floor.
DR
ROBERTS:          We are
in agreement on that.  I mean, it wouldn’t necessarily

sell all of its output, but I that’s … I think we
agree.’
[95]
By contrast, the Commission insisted that
FAV must determine the cost of feedstock propylene.   In
support, it referred
to the evidence of Dr Koster, on behalf of
appellant, which was to the effect that US prices were below FAV
(being alkylation values)
in 2006 and at FAV in 2008.  This set
of figures accorded in the Commission’s view of the way in
which the market works;
that is in a competitive market prices are
competed down towards cost.  If the market for the supply of the
input becomes
competitive (with many firms with similar cost
structures), the market price of the input will fall towards cost
because the sellers
have no bargaining power.
[96]
In support of these arguments the
Commission referred to Sasol’s own contemporaneous internal
documents which reflected that
FAV was a market price including the
1995 transfer pricing policy and the 2001 transfer pricing policy.
These prices were
reflected in Sasol’s annual financial
statements which were representations intended to be true and
reliable made inter alia
to shareholders investing public to JSE and
the SEC in the United States.
[97]
Mr Wesley submitted that, in defining the
economic value of feedstock propylene at appellant’s breakeven
price, which is the
highest price that Synfuels could charge, by
definition this would lead to the conclusion that appellant’s
prices were not
excessive because they were not higher than its
costs.  This would move the excessive pricing upstream while
preventing scrutiny
of the upstream producer, being Synfuels, because
it is not actually charging that feedstock price in that the price
being employed
would be a constructive price.
[98]
It is now possible to return to the
relevant jurisprudence.  In the
Mittal,
supra
case this Court sought to give
content to the definition of excessive price as contained in s 1 (1)
(ix); that is a price for a
good or service which (aa) bears no
reasonable relation to the economic value of that good or service;
and (bb) is higher than
the value referred to in subparagraph (aa).
As noted in the introduction to this judgment, this definition needs
to be read
together with s 8 (a) which provides that it is prohibited
for a dominant firm to charge an excessive price to the detriment of

consumers.
[99]
The analysis with which this judgment has
been concerned has turned on the question of an excessive price, as
that term has been
interpreted.  The interpretation seized upon
by the Tribunal in much of its determination was informed by a
reading of but
a few paragraphs of the
Mittal
,
judgment.    In dealing with this question at para 43
this Court said:
‘
Thus
economic value is a notional objective competitive market standard,
not one derived from circumstances peculiar to the particular
firm.
If the firms prices are no higher than economic value, no
contravention of s 8 (a) can arise.   If
however, the
firm’s price is in fact higher than economic value so
determined, the test of reasonableness in respect of the
difference
remains to be applied… the test of reasonableness applies to
the excess of price over economic value and thus
only to the element
of ‘pure profit’ (over and above normal profit) implicit
in that price.’
The
court went on to say:  ‘
The
criterion of economic value … recognises only the costs that
would be recovered in long run competitive equilibrium.’
It also said in para 42 that at the stage of
reasonableness, account could be taken of benefits flowing to the
firm ‘
from the subsidised loan, long
term low rental or other special advantage which may serve to reduce
its own long run average cost
below the notional norm.
’
[100]
Unfortunately, both during the evidence
that was led and in the determination by the Tribunal, only these
paragraphs from the judgment
of this Court were canvassed in any
detail.  What renders the Tribunal determination even more
confusing is that, after a
lengthy exposition of the law and
application to the facts, all based on these paragraphs, the Tribunal
belatedly refers to later
passages of the judgment. This piecemeal
reading is regrettable.
[101]
I shall return
presently to the role of an expert in these matters but, at this
stage, suffice it to say that Dr Roberts, who was
called to give
expert evidence on economic questions, showed an unfortunate keenness
to provide the Tribunal with the benefit of
his legal expertise and
his own reading of the judgment, notwithstanding that he has no legal
expertise.  The Tribunal should
guard against the unfortunate
tendency in its future hearings.
[102]
This unsatisfactory, partial reading of the
Mittal
judgment only compounded the difficulties experienced by this Court
on appeal.  A full reading of the
Mittal
judgment shows that the Court engaged fully with the effect of the
United Brands
decision,
supra
which so heavily influenced the formulation of s 8 (a) of the Act.
Of particular relevance is the following passage at paras
49-50:
‘
A
‘fairly robust approach’ may thus have to be adopted
particularly when account is taken that ‘long run normal’

profit and the conceptual basis upon which this term is predicated
are notional.  Within the context of adjudication, which
deals
with probabilities, these concepts cannot be employed with scientific
precision.  For example, where the actual price
is shown, as in
the
British Leyland
case, to exceed the normal price for roughly similar products to a
degree which is, on the face of it, utterly exorbitant, then
the need
to quantify economic value more precisely before concluding that the
actual price bears no reasonable relation to it may
fall away.
In
this way a
prima facie
case would have been made out, leaving
it to a firm in appellant’s position to adduce evidence to the
contrary.  If
it is to avoid the case against it becoming
conclusive.  Likewise, where the dominant firm raises the normal
price for its
product substantially without any corresponding rise in
costs, this may indicate
prima facie
that the new price is
higher than economic value without the need to quantify the latter
more precisely.  Where input costs
vary considerably in cycles,
the dominant firm’s actual costs may fall sharply without it
carrying out a corresponding reduction
in its price.  Likewise,
if the firm usually prices to import parity, it may neglect for a
time to bring its price into correspondence
with that (ultimately
constraining) maximum, relying in the short term on customer
ignorance or inertia in order to charge more.
In consequence,
the firm’s own accounting profits may show a considerable
increase during a certain period or periods, over
and above the
levels which it usually achieves.  If there is no reason to
suppose that the firm’s own usual levels of
accounting profits
would have resulted in a return on capital that is less than the
notional competitive norm, (ie enough to sustain
it in business in
the long run), then it would appear
prima facie
that the firm
must have earned ‘pure’ profit as a result of its pricing
during the period or periods when the spike
occurred.  Thus an
adverse finding on a narrower basis that that originally alleged may
potentially be secured, without any
concession that the firm’s
prices ordinarily charged when input costs etc were higher were
themselves legitimate.’
[103]
Given the influence of European law, the
decision in
Scandlines Sverige AJ v
Helsingborg
(unreported decision of the
European Commission Case No: COMP/A 36.568/D3) is also of relevance
to the explication  of the
phrase “excessive value”.
[104]
The following passages at paras 102-103 of
that decision are of particularly interesting:
‘
It
is important to note that the decisive test in United Brands focuses
on the price charged, and its relation to the economic value
of the
product.  While a comparison of prices and costs, which reveals
the profit margin, of a particular company may serve
as a first step
in the analysis (if at all possible to calculate), this in itself
cannot be conclusive as regards the existence
of an abuse under
Article 82.
In
this decision, the Commission will follow the methodology set out by
the Court in paragraph 252 of the United Brands judgment.
The
Commission will therefore assess the costs actually incurred by HHAB
in providing the products/services in question (the costs
of
production) and make a comparison with the prices actually charged
(section II.B.2.1).   The Commission will then
assess
whether the prices are unfair when compared to prices charged to
other users or by other posts (section II.B.2.2), or whether
the
prices are unfair in themselves (section II.B.2.3).’
[105]
The costs which were incurred for the
feedstock propylene reflected the price charged by Synfuels which
became known as the transfer
price.   This was the price at
which Synfuels sold feedstock to appellant.   This was the
price at which appellant
acquired the feedstock.  Dr Padilla,
obviously aware that at this level of costs, appellant ran the risk
of being found guilty
of s 8 (a) as a result of possible significant
differences between price and costs sought to construct a cost based
basis on the
notional idea of a long run competitive equilibrium at
which various suppliers could provide feedstock but only Synfuels
would
have the advantage of a low cost structure which it had to pass
on to appellant.   By contrast, Dr Roberts sought to reduce

the costs below that which Synfuels charged appellant in order to
strengthen the case which the Commission sought to bring before
this
Court.
[106]
This almost surrealistic commitment to
either a high cost, being the refineries price of feedstock, or a low
cost based on reductions
from the transfer price were luminously
illustrated in certain passages of the evidence.  For example
under cross examination
Professor Wainer, the accounting expert who
testified on behalf of the Commission, was pressed by Mr Trengove
with regard to the
appropriate costing of feedstock:
‘
ADV
TRENGOVE: What do you say about feedstock?
PROF
WAINER: I said I don’t express a view as to whether it is…
was sold at below or above market value. I simply made
the point that
that these runs contrary to the disclosures in the financial
statements.
ADV
TRENGOVE: And the same could be said of Dr Roberts’ adjustment
of feedstock costs?
PROF
WAINER: Yes.
ADV
TRENGOVE: You also said that if SCI were…
PROF
WAINER: Can I put this judgment away?
ADV
TRENGOVE: For the time being, yes.
PROF
WAINER: Hopefully forever.
ADV
TRENGOVE: Well I don’t make any promises. You also say that if
SCI were earning the returns indicated by Mr Harman’s

calculations.
PROF
WAINER: Yes.
ADV
TRENGOVE: Then it should close its business.
PROF
WAINER: Yes.
ADV
TRENGOVE: But that is because you equate economic cost with
accounting costs, is that not so?
PROF
WAINER: No.  If you were to measure the returns in this fashion,
forget about whether it’s accounting or economic,
I’m
talking about in a financial context, in other words, if this was a
representation of reality, then if that’s the
reality when you
must close shop, because you’re generating returns so far below
your WACC or below any reasonable rate of
return.
ADV
TRENGOVE:
But nobody has suggested that
it reflects a reality at all
.
PROF
WAINER: But that’s the implications I’ve simply pointed
out.  I’m not
saying that
they suggested it was reality
.
What I’m pointing out to the extent that it may be useful, is
what a peculiar result is achieved, that if you were
to rely on this
as being any representation of something of value, then the necessary
inference is that they should close shop.’
(my emphasis)
[107]
But it was not only under cross examination
that the dichotomy between ‘a reality’ and a costing
based on a particular
form of economic modelling was reflected.
Mr Behrens who testified on behalf of appellant was cross examined on
a similar
topic.  The following passage of evidence is equally
illustrative.
‘
MR
BEHRENS: In a competition case you need the expert analysis of Mr
Harman. However, what I can say that in the business we know
that by
just, call it making a profit on the accounts is not necessarily
sufficient to say that it is economically profitable.
ADV
WESLEY: No, but the extent to which you make money over operating
cost or cash cost, the reward you need, what is the proper
measure of
cost, do you use the replacement costs of the asset, you will recall
we have spent days on this with the expert evidence.
You recall
all of that?
MR
BEHRENS: Yes.
ADV
WESLEY: Yes and so, your statement that currently Sasol polymers is
not economically profitable in that sense would be a matter
of expert
opinion, because you would have to have the similar debates about
what costs are you looking at, what is the regard you
are looking at,
how do you measure all of these things.  You accept that?
MR
BEHRENS: What I am leading to in my answer is that within this
connect, yes, within this case, but within running the business
we
would always bear in mind the aspects that I am mentioning.
Those are real aspects like what is the working capital that
is tied
up in that company.  So, even if it might be making a profit, if
it is tying up a lot of working capital, management
will star looking
at the company and to say well even then it might be the right thing
to shut the business, because it is not
rewarding the assets and the
money that it is holding up or that it requires to operate.’
[108]
The uncontested fact reveals that the
actual price paid by appellant during the entire complaint period was
the price as reflected
in the 2003 One Tier agreement.  Mr
Behrens said the following about the price contained in that
agreement:
‘
In
the case of SP’s Monomers business, however, Synfuels and SP
belong to the same group of companies.  This has resulted
in
supply agreements in which Monomers (and previously Polifin) has
obtained feedstock effectively at Synfuels’ (comparatively
low)
FAV.  Even if a notional propylene manufacturer in South Africa
could buy propylene feedstock from Synfuels, the latter
would not
supply an independent manufacturer on the advantageous terms that SP
enjoys.  It is naturally a speculative exercise
to say what
price Synfuels would obtain from an independent purchaser in this
hypothetical scenario, but it is certain that Synfuels
would not
allow the purchaser to reap all the benefit of the low FAV of
Synfuels’ propylene feedstock.  The price would
certainly
be higher than Synfuels’ FAV and probably higher than an oil
refiner’s FAV, since Synfuels (like feedstock
suppliers
globally)  would seek a price which would ensure that the
propylene purifier could obtain no more than a market
related margin
in on-selling polymer-grade propylene to its customers.’
[109]
That a price advantaged a company within
the group is, as Mr Behrens suggests, unsurprising.  But the
price was pitched at
a level which would pass transfer pricing
legislation.  This is evident in that the relevant financial
statements were drafted
to pass tax muster, which means that they
were required to comply with an ‘arm’s length
standard’.
Accordingly, there appears to be an
insufficient basis to intervene in the complex problem of pricing.
A measure of deference
is called for in these enquiries not only
because of the importance of freedom of pricing but also to obviate
converting courts
into price regulations.
[110]
Two further observations need to be made.
As the European Commission noted in
Scandlines
Sverige, supra
at para 232:
‘
The
economic value of the product/service cannot simple be determined by
adding to the approximate costs incurred in the provision
of this
product/service … a profit margin which would be a
pre-determined percentage of the product costs.   [Rather,

the] economic value must be determined with regards to the particular
circumstances of the case and take into account also non-costs

factors such as the demand for the product/service.’
[111]
In the
Mittal
case, this Court was concerned to deal with the pricing policy
adopted by
Mittal
.
In this case, the Tribunal had, in essence, taken the view that, once
a firm is super dominant, a price that cannot be found
to be ‘based
on cognisable competition considerations is excessive, in that it
will not have been determined by the free
interaction of demand and
supply in a competitive market.  Faced with a Tribunal decision
that totally ignored all the detailed
evidence led before it, this
Court in
Mittal
sought to provide a framework to evaluate this evidence and thereby
determine whether the price so charged was excessive; hence
the
importance of para 40 and 43 of the judgment.  In the present
case, the key question turned on a different issue: the
refusal to
pass on a cost advantage which turned not on the pricing policy of
appellant alone but also on Synfuels, which was not
a party to these
proceedings.  In this case therefore, if the cost of an
essential component of the production of product/s,
whose prices are
under scrutiny, can be justified on rational grounds, that should be
the yardstick employed in the primary inquiry
with which the Court is
engaged.   The complexity of price assessment dictates that
some deference is required
[112]
It is possible to illustrate
the implications of this finding with the following hypothetical.
Chemical X is very volatile.
It is neither imported nor
exported.  In Korea, United States, China, Europe, the
conventional feedstock import for
producing X is extracted from deep
mines at a costs of R 100.   There are additional
production costs of R 20, resulting
in total costs per unit of R
120.  These markets are highly competitive and the price of X is
approximately R 125.  In
South Africa, by contrast, the
feedstock lies just below the surface and the cost of this feedstock
is low, being only R 10.
All the land where the feedstock
resides is owned by A Co.  A Co’s costs of producing X is
only R 30.  A Co sells
X in South Africa at a price of R 100
which is well below the price of other countries, where the
production costs are R 120 in
without taking into account the
competitive costs of capital.  A Co would argue that the price
is not excessive in relation
to some international price benchmark
which is higher.  A Co says that R 100 is fair because it prices
lower than the prices
of other countries.   But it is
excessive in relation to A Co’s costs.  When it is
recalled that s 8(a) is
concerned to deal with a dominant firm
(during the complaint period, appellant’s national market share
for the production
and supply for purified propylene measured by
capacity was in excess of 90%) then, in effect, the monopolist’s
costs become
relevant.  Suppose it is possible to import X into
South Africa at transport costs of R 10.   The import
‘landing
costs’ in South Africa would now be R 130.
Suppose the opportunity costs of importers selling into South Africa,
that
is taking into account the fact that importers could sell in
their own countries, is R 135.   Suppose that A Co now
argues
that the price of R 100 is not excessive because it is less
than the price of substitutes which is R 135.   The
argument
is that R 135 is a competitive price, that is the price of
the next best substitute.   But A Co still earns profits
which
approximate those of a monopolist.   The monopoly
price is defined by the price where the monopolist maximises the
profits,
that is where it is unprofitable to raise the price by even
more.  A Co lacks the incentive to raise the price above R 100

because the price of R 100 maximises profits.  It supplies the
whole market therefore at R100.  To use the cost of substitutes

in the way X suggest would be to declare that no firm would ever be
seen as pricing excessively; that is there is always some next
best
substitute, just one that is far less economic or practical.
[113]
This scenario, sketched by way of a
hypothetical, is not far removed from the well-known American case of
The United States v Aluminium Company of
America
148 F. 2d 416
(2d Cir . 1945),
where Judge Learned Hand was required to analyse a situation where
Alcoa priced above its costs but below the
landed costs of imports
which were higher because of the transportation cost and the
imposition of a tariff.  Judge Hand said:
‘
It
is entirely consistent with the evidence that it was the threat of
greater foreign imports which kept Alcoa’s price where
they
were, and prevented it from exploiting it advantage as sole domestic
producer; indeed it is hard to resist the conclusion
that potential
imports did put a “ceiling” upon those prices.
Nevertheless within the limits afforded by the
tariff and the costs
of transportation, Alcoa is free to raise its prices as it chose
since it was free from domestic competition
save as it drew other
metals into the  markets as substitutes.’
[114]
Assume a slightly different scenario.
A Co sells at a price of R 134 barely below the constraint
implied by the potential
imports at R 135.  The argument is that
this is a competitive price for it is R 1 lower than any other
competitive price by
way of an import.   The invitation by
A is in effect for a court to fall for the so called cellophane
fallacy (
United States v EI Dupont
Devemours and Company
[1956] USSC 87
;
351 US 377
(1956).  Richard Posner
Antitrust
Law
(2001) at 150-151 captures the
fallacy thus:
‘
If
a firm has monopoly power, it is probably already pricing at a point
reflecting the use of its monopoly power, and just below
the level
that will cause a critical mass of buyers to shift to substitutes.
At that point, by definition, alternative
products and their sellers
will keep the monopolist’s  price at its current price and
no higher.  So one would
expect an increase in a profit
maximising monopolist’s price to cause buyers to shift to an
alternative product.  Therefore
cross-elasticity of demand for a
product calculated on the basis of current price tells you only the
outer limits of the power
of a putative monopolist.’
[115]
In the light of this analysis, the ultimate
determination of whether appellant has contravened s 8(a) must be
predicated on a feedstock
price which represents the price which
Synfuels sold feedstock to appellant. That price can be justified as
the actual cost pursuant
to an agreement that is subject to
independent regulatory scrutiny. There is no basis to adopt a
hypothetical price, divorced from
the reality conceded by appellant,
which application would gut any possibility of prosecuting on
excessive pricing case.
[116]
The balance of the illustration and the
hypothetical remains to be determined: that is whether, given the
returns enjoyed by the
appellant, the price for the products in the
defined markets in South Africa, constituted a price which is
excessive in terms of
economic value.   Before we arrive at
this conclusion however, the remaining costing issues require
attention.
Evaluation
of capital assets
[117]
An evaluation of appellant’s capital
assets is relevant to the ultimate determination of the costs
incurred in the production
of the relevant products for the following
reasons.
1.
There is a dispute concerning the costs of depreciation; that is the
annual cost of the capital assets employed in appellant’s

business
2.
There is a dispute concerning the return to investors; that is the
reward to investors for their investment in appellant.
On behalf of the
Commission, Dr Roberts valued the capital assets at their book value;
that is he assessed the historical costs
of the assets less
depreciation at the beginning of the complaint period.  The
Tribunal preferred the approach of appellant’s
expert, Mr
Harman, who employed a replacement cost value.  This approach
required a revaluation of appellant’s capital
assets by
updating the historical costs for inflation by an industry inflation
index which is designed to keep track with inflationary
increases in
plant costs.  On the basis of this approach, the depreciated
book value is inflated in order to determine the
value of second hand
assets at current prices at the beginning of the complaint period.
In his evidence Mr Harman provided
a simple example to illustrate the
obvious attraction of his approach;
‘
MR
HARMAN: So just give you an example of that, let’s say that I
purchased an asset some time ago for R 100.00, it has been

depreciated to today, so that it’s worth R 50.00.  Now
let’s say that there had been 100% inflation over the period,

all that I’m doing is I’m taking that 50 and inflation
that number up by inflation.
PROF
HOLDEN: Inflation of what?
MR
HARMAN: The inflation of the asset by an HIS CERA index.
PROF
HOLDEN: And how did you get that?
MR
HARMAN: The inflation index?
PROF
HOLDEN: Yes.
MR
HARMAN: The inflation index comes from CMAL, or it may have come from
SCI, I can’t remember the ultimate source, but it
is an asset
index that is used widely in the chemical industry to trace how
chemical plants and maybe other plants increase over
time.
So it’s an industrial index, not a CPI index, correct.
ADV
TRENGOVE: An industrial index designed for the very purpose of
keeping track with the increases in plant costs?
MR
HARMAN: Correct and so my use of this has been confirmed by CMAI as
being a sensible way to increase assets, but I haven’t
left it
there.   So that is my base and I’ve tried to select
an inflation profile that gives me a low value.’
[118]
By contrast, the Commission contended that
appellant’s assets were in the early part of their life and so
the reward on them
was higher than would be the case if the
calculation was done on an average return over their use for life.
The Commission’s
approach employed a reward on capital invested
at historical costs which took into account that, if the assets were
significantly
depreciated, there would be a relatively lower return.
Further, the Commission considered that Mr Harman had approached the

matter on an incorrect premise; that is that appellant’s assets
would need to be replaced.   In the view of the
Commission,
this would not make sense in this market and industry and did not
recognise the outcome under effective competition
between established
rivals.  Rivalry between existing companies would mean rewarding
capital invested for incremental expansions
as well as maintaining
the plant.  The main assets in this industry generally had very
long lives.
[119]
The problem for the Commission was that
their own expert, Professor Wainer, conceded that the historical
costs basis for depreciation
provides only for the replacement of
assets at the end of its life at the amount of the original
historical cost.  This calculation
makes no provision for the
effect of inflation, because it values assets at the price at which
they were originally purchased.
Manifestly, the logic
employed by Mr Harman within the context of an inflationary
environment is a more logical and coherent approach
to the problem of
depreciation. This approach was captured in
Scandlines
Sverige AV, supra
at para 223 –
224 where the European Commission made the following point:
‘
However
a company that sets its prices on basis of depreciated historical
costs may – depending on how the production costs
of the
relevant assets have developed over the years – will find
itself in the position that its return does not (ie. No
longer) allow
it to finance future capital expenditures for the replacement of
existing assets.’
[120]
The Tribunal however rejected Mr Harman’s
contention for an additional reduction in this regard, by using
insurance values
as the appropriate value of replacement costs:
‘
The
insurance values of a firm’s assets may for any number of
reasons would be either over or understated.  If one is
going to
value assets by reference to an insurance value one needs to
interrogate the insurance value methodology because insurance

standards and economic standard may all differ.   This was
not done.’
[121]
In this regard, the Tribunal did not appear
to base its adverse finding on any significant argument raised by the
Commission.
Indeed, in the expert minute of 10 May 2013, the
only objection taken to insurance values was it was ‘inappropriate
–
as form of replacement costs”.   No further
justification was offered.  In his supplementary report of 30
July 2012 Mr Harman justified the use of insurance values as follows:
‘
I
consider that this is a reasonable assumption because: i) the
insurance values are accepted by Sasol’s insurers; ii)
insurance
values do not include all of the assets associated with
SCI’s facilities (i.e. they exclude many of the commercial
assets,
such as buildings, workshops, IT equipment and
infrastructure); and iii) insurance values are lower than CMAI
estimates based on
a Greenfield basis.  For the financial year
2001 to financial year 2009 period, I have used an insurance value of
R573m on
a net replacement cost basis.’
[122]
There was no answer from the Commission to
this compelling argument nor any reasoning employed by the Tribunal
which would justify
a rejection of Mr Harman’s approach.
Capital
Reward
[123]
It was common cause that the cost of equity
capital must factor in a return on the capital employed.  Two
disputes arose in
this connection:
1.
How to determine the equity risk premium.
2.
Whether the benchmark is a pre-tax or after tax measure.
[124]
Mr Harman based his determination of the
normal return on capital on a computation of the weighted average
costs of capital (‘WACC’)
of appellant and, within the
determination of the WACC, employed the capital asset pricing model
(CAPM) for the determination of
the cost of equity.  In his
first report together with his third report, Mr Harman noted that his
calculations were based
on ‘the analytical framework of the
WACC which assumes that companies are financed by a mix of debt and
equity capital.
Both forms of finance have different costs
associated with them and the WACC reflects the average return that an
entity must achieve
to satisfy the demands of both equity and debt
holders’.
[125]
Mr Harman calculated the costs of debt by
employing the interest rate paid on South African government debt and
adding a premium
to reflect the additional risk that debt holders
face when lending to Sasol rather than to the South African
government, a point
which hardly requires justification.
In his calculation of the costs of equity use Mr Harman employed the
CAPM which
is widely used by South African companies for the
valuation of their equity investments as well as by South African
regulators
and regulators globally.
[126]
Although there is a debate relating to how
CAPM can be calculated, Mr Harman followed standard practice,
providing a range of estimates
for an average  WACC which
estimates reflect the inherent uncertainty in calculating the cost of
capital.   In validating
these results he compared them to
a number of external sources.
[127]
By contrast, Professor Weiner testified
that the normal return on equity was a risk free rate together with
an equity risk premium
of between 4-7%, Dr Roberts adopted a risk
free rate together with 3% as his basis for evaluation.
For reasons which
I propose to set out later in this judgment, both
the evidence of Professor Weiner and Dr Roberts in this regard leaves
much to
be desired.  These calculations were no more than a
‘thumb suck’.  Unfortunately, this Court was thus
disadvantaged
by not benefitting from evidence on this point from the
Commission that was adequately reasoned.
[128]
The Tribunal was therefore clearly correct
in dismissing this evidence.  However, for reasons that are
hardly justified and,
to the extent that there was any justification
little sense can be divined therefrom, the Tribunal rejected Mr
Harman’s proposed
use of an inception of WACC.  Having
accepted his method and thus logical approach adopted by Mr Harman,
the Tribunal should
have accepted appellant’s case which was
based on the use of an average inception WACC over the complaint
period.  That
is not to say that the use of the WACC and hence
the CAPM is beyond any critical reproach.  See for example
Richard Rolls
1977
Journal of Financial
Economics
129 and Richard Dayala 2012
(31)
Business Valuation Review
23.  By contrast see the Interpretations Committee of the
International Accounting Standards Board’s Guidelines on the

Application of IFRS 13 Fair Value Measurement December 2012.
But no justifiable evidence to counter Mr Harman’s testimony

was offered by the Commission.
[129]
In his evidence, which was neither
contested nor adequately rejected by way of a set of plausible
reasons offered by the Tribunal,
Mr Harman testified that, to
preserve an investment incentive, a company must be rewarded for the
risk expected ex ante.   Therefore
setting an ex ante cost
of capital, that is the costs of capital used in assessing the
project at its inception, is a necessary
component of a coherent
analysis.  In his expert report Mr Harman explains this as
follows:
‘
I
have carried out a number of profitability tests.  It can be
argued that, in assessment of realised rates of return, the
relevant
costs of capital should be the costs of capital that was used in
assessing the project at its inception.  In this
case, the
relevant costs of capital would be SCI’s estimated “project
inception” WACC in 1990.   However,
I have considered
the following project inception costs of capital:
·
the cost of capital for financial year 2000
as an indicator of the project inception WACC, which I compare to
profitability over
the period 2001 to 2008; and
·
the cost of capital for financial year 2003
as an indicator of the project inception WACC, which I compare to
profitability over
the Complaint Period.
As
the risk-free rate in South Africa dropped significantly over the
period 1990 to 2001, my estimates of the project inception
WACC are
lower than the WACC in 1990.  As such, my estimate of the WACC
at project inception will be conservative.’
[130]
There was also no plausible reason
for rejecting the use of a hurdle rate which clearly was necessary
because appellant sought to
price its projects on an ex ante basis.
The hurdle rate allows for an estimation of specific risks as
contrasted to an ex
post facto outcome that does not account for
these risks.
[131]
The second dispute concerned a before or
after tax return.   Regrettably, the evidence of the
Commission again reflected
a level of analysis which could not
possibly be plausibly advanced by an expert in the field.  In
the assessment of a marginal
investment, a firm would take account of
the marginal effective tax rate, which is a measure of the burden of
tax on the investment
for a profit maximising firm. This rate
determines the scale of a project: a higher marginal effective tax
rate means small size
projects and fewer investments.  The
marginal effective tax rate is well known as an important
determination for foreign direct
investment and shows manifestly
that, when examining returns, it is the after tax return upon which
an investor relies in order
to assess the possibility of a new
investment.
Group
Costs
[132]
Regrettably there is little reasoning in
the approach adopted by the Tribunal with regard to this issue.
Three sets
of costs were involved, corporate development costs,
management fee costs and insurance costs.   In seeking to
divine
the reasoning by which the Tribunal considered that no
adjustment for group costs was to be made, the only plausible reason
that
I could find is contained in the following two paragraphs of the
Tribunal’s determination.
‘
The
evidence suggests that the corporate development costs, which
represents approximately 65% of the group costs allocated to
propylene, were actually for the Fischer Tropsch technology and the
associated research.  We have found no direct link between
the
supposed value of this research to purified propylene and the costs
allocated to purified propylene.  There was also no
evidence
that the costs allocated to propylene were incurred for propylene.
Furthermore,
Behrens gave evidence that the group costs were allocated to Sasol
Polymers on an operating margin basis.  However,
this would give
rise to a bias as more costs would be allocated to the businesses
that have high operating margins rather than
the businesses that
realise the most value from the research.’
[133]
Mr Harman  had prepared a report which
was based on appellant’s discovered information and which
revealed that in 2009
PP was allocated R 101.3 million of total group
costs and propylene R 56.3 million.  According to Mr Harman,
over 80% of these
amounts were made up of three costs items:
corporate development costs, Sasol management fee costs and insurance
costs.  Mr
Harman had taken these 2009 and the allocations and
deflated them for inflation.   He argued, on this basis,
that the
cost allocation method that he had employed was reasonable
and reflected, in broad terms, the group costs that would have been
allocated to propylene and PP during the complaint period after
adjusting for the impact for inflation.   The submissions

made by the appellant that the group costs allocation applied by Mr
Harman reflects a reasonable estimation of the group costs
incurred
for the benefit of propylene and PP during the complaint period and
should be upheld.
Allocation
of common costs
[134]
The Tribunal adopted a volume based
approach which had been proposed by the Commission with regard to the
allocation of common costs
for propylene.   The appellant
accepted this allocation methodology for, in its view, it had a
trivial effect on the
results.   With regards to PP
however, the appellant rejected the approach adopted by both Dr
Roberts and Professor Weiner
that a volume based allocation to costs
between export and domestic PP sales was appropriate.  These
witnesses assumed that
the fixed costs were incurred equally across
all products, an assumption favoured by the Tribunal, on the basis
that the allocation,
having been made related to the same product.
Further it considered that the costs of production were the same of
that product
irrespective of the location of the sale.
[135]
Mr Harman offered a most plausible rebuttal
of this approach which is reflected in his evidence as follows:
‘
If
costs are allocated on a volume basis, [PP] exports have a price-cost
mark up of minus 12.5% and domestic sales have a price-cost
mark up
of 2.4%.  Over both markets the average price-cost mark up is
minus 5.6%.  This suggests that exports are loss-making,
whereas
domestic sales recover economic costs.  However, the negative
price-cost mark up in the export market suggests that
Sasol
Polypropylene would be better off ceasing production for export
markets. Ceasing export sales would leave Sasol Polypropylene
with
only domestic sales but its fixed costs would remain unchanged.
This would leave a price-costs mark up of minus 15.9%.

Eliminating the apparently loss-making exports thus leads to a
reduction in average profitability from minus 5.6% to minus 15.9%.

This counter-intuitive result reflects the fact that exports only
appeared loss making because the volume base allocation over

allocates common costs to exports.’
[136]
Although Mr Harman conceded that there was
no one correct method of allocating costs to a single product with
different prices and
different investment economics, he indicated a
strong preference for an allocation in which common costs were
allocated to the
export business up to the point where it recovered
economic costs.  The remainder would then be allocated to the
domestic
business.  This approach recognises the marginal nature
of export sales by allocating proportional economic costs to the
export
business which breaks even in economic terms.  The
remaining fixed costs are then paid by the domestic business.
Any
greater allocation of common costs to the export business would
suggest that appellant’s export at a loss which it did not
do.
[137]
I accept the approach adopted by the
appellant was the most rational approach provided towards the
question of allocated costs.
Aggregate
of separate calculations for Tier 1 and Tier 2 prices
[138]
The supply agreement concluded between
appellant and Safripol during the complaint period (following an
amendment of 26 June 1999
was that Safripol would pay appellant a
Tier 1 price for the first 55 000 tons propylene purchased by
Safripol.  The
Tier 1 price was “calculated on a monthly
basis i.e divided by 12”.   Thereafter the Tier 2
price would apply.
In his evidence Mr Behrens explained the
structure as follows:
‘
The
increased price charged by SP to Safripol for volumes above 55 000
tons was directly related to higher costs associated
with the
production of the additional volumes, including a higher feedstock
price from Synfuels (as just explained), higher capital
costs
incurred by SP in constructing PPU3 as compared with PPU1 and higher
operating costs due to the dilute nature of the relevant
condensate
stream.
It
is important to note that the increased price paid by Safripol for
volumes above 55 000 tons was also applicable to SP’s
PP
business for its additional volumes arising from the new investment
(PPU3).’
[139]
Safripol’s monthly bill therefore
comprised a portion of its supplies for the month at the Tier 1 price
and the balance at
the Tier 2 price.  Each monthly bill
reflected an amount equal to the weighted average of the two prices.
Appellant contends
that the propylene prices which should be
considered for the purposes of assessing whether or not its prices
were excessive is
the weighted average of appellant’s propylene
prices for propylene sold to Safripol.
[140]
By contrast, the Tribunal found that the
Tier 1 and the Tier 2 prices should be considered separately and
hence it would be inappropriate
to consider a weighted average.
The Tribunal found that the evidence suggested that Safripol’s
decisions to buy were
made on the Tier 2 prices independently of the
Tier 1 prices.   Safripol bought all of the contractually
fixed volumes
of purified propylene at the lower Tier 1 price but did
not purchase all of the volume it could at the higher Tier 2 price.
[141]
The Tribunal noted that appellant contended
that the price differentiation between the two Tiers reflected higher
costs associated
with the production of the additional purified
propylene volumes, including a higher feedstock price from Synfuels,
higher purification
and operating costs due to the diluted nature of
the relevant condensate stream and necessary investments.
However
for the Tribunal, these allegations did not accord with the
way in which the propylene pricing was negotiated with Polifin nor
did it accord with the interpretation provided by Mr MacDougal who
explained it simply as a ‘take it or leave it’ offer.
[142]
I agree with Mr Trengove when he contended
that Safripol paid the weighted average of the Tier 1 and Tier 2
prices every month during
the complaint period.  Accordingly,
the “real” price paid by Safripol every month was the
weighted average of
the two prices.  It appeared to be a rather
artificial construct to separate them on expost facto basis.
Significantly,
when the costs of Tier 1 and tier 2 were examined, the
Commission did not separately deal with the costs of Tier 1 and Tier
2 but
compared Tier 2 prices to average costs, which approval was
followed by the Tribunal in its calculations.
[143]
To the extent that it was suggested that
there was no evidence to show that there was difference in costs
associated with the production
of Tier 1 and Tier 2 propylene, as I
have indicated Mr Behrens provided an explanation.   No
further evidence on either
side was forthcoming.
[144]
However the following cross examination of
Dr Roberts is also instructive:
‘
ADV
TRENGOVE: So, is although they had this formula with two components,
the monthly bill would always be a bill in effect a weighted
average
of the two prices.
DR
ROBERTS: Ja, the bill would be on an average price. The decision by
the firm, of course, is completely different, because the
buyer is
looking at two separate prices, but I accept that that’s the
way in which the bill will be sent.’
[145]
In attempting to fend off this concession
by Dr Roberts, Mr Wesley contended that while Safripol does buy all
of the propylene volumes
at the lower Tier 1 price, it does not buy
all the volume it could have at the Tier 2 price, because it buys
further product from
Sapref.   I fail to understand how
this additional purchase trumps the argument that a calculation which
sets out the
average price should not be the yardstick for the
overall assessment.
[146]
The Tribunal found it would be
inappropriate to consider the weighted average mark up of propylene /
PP over the costs of the complaint
period.  It simply stated
that ‘each year should be attributed equal weight’ but,
as unfortunately was the case
in much of this determination, it never
properly explained why this attribution should apply in this case.
[147]
By contrast, Mr Harman gave credible
evidence which is summarised in his third report as follows:
‘
When
Dr Roberts summarises his multi-year analysis into a single
percentage, his average is based on the average mark up in each
year
(i.e is a simple average).  I calculate a percentage based on
the total costs and revenues over the period (I.e a weighted

average).  This difference in arithmetic approach creates a
difference of approximately 2% between our results for both propylene

and polypropylene.  I consider that my approach avoids a
calculation bias.’
Mr
Harman expanded on this evidence as follows:
‘
My
approach avoids an upwards bias Dr. Roberts’ approach.
Consider a product with a constant unit price of 100 for two
years
but unit costs of 80 in Year 1 and 120 in Year 2.  The product
gains 20 (100 minus 80) in Year 1 but under-recovers
20 (100 minus
120) in Year 2.  My approach would find a zero percentage mark
up (i.e sum of total revenue (100+100=200) divided
by total cost
(80+120=200) minus 1) Dr Roberts’ approach, would find a
positive mark up of 4%.  In Year 1, the mark
up would be 25%
(100/80-1) and in Year 2, it would be minus 17% (100/120-1).
The average of 25% and (17%) equals 4%.
My approach is
consistent with IRR approach, which considers total revenues and
total costs over time rather than revenues and
costs for each
individual year in isolation.’
[148]
Given that this case is about a complaint
period and the establishment of excessive pricing over this
particular period, and not
in specific years of the complaint period,
to employ a simple average as was applied by the Commission would
have a computational
upward price whereby prices that exactly equal
average economic value should be found to exceed this value.
Mr Harman’s
evidence should have been favoured and accordingly
a weighted average should have been adopted by the Tribunal.
[149]
To recapitulate this Court has found that
the feedstock propylene costs should be held to be equal to the price
charged by Sasol
Synfuels in terms of the 2003 agreement.
It has also found that the approach adopted by the appellant to the
evaluation
of appellant’s capital assets, the appropriate rate
of return on capital, the allocation of group costs and the
allocation
of common costs should be followed.
[150]
The Court then found itself in a
difficulty. No calculation had been made by either of the parties of
the price – cost mark
up on the basis of these findings.
Accordingly, in a series of requests to the parties on 17, 20 and 23
April 2015 this Court sought
assistance with regard to these
calculations. For completeness I should add that in its second letter
of 20 April 2015, the Court
clarified its request, requesting that
the parties conduct the calculation on the following basis:
1.
The Tier 1 and Tier 2 Propylene prices are treated separately; and
2.
On an averaging of Tier 1 and Tier 2 purified propylene prices.
[151]
On 23 April 2015 a further request was
generated which confirmed that the following assumption should also
apply:
1.
The capital assets should be valued at their depreciated insurance
values.
2.
The inception WACC including a hurdle rate should be used to
determine the capital reward charge.
3.
Common costs should be allocated using the volume formula for
propylene and the economic method for PP.
4.
The analysis should be performed over two periods from FY 2001 to FY
2008 inclusive and from FY 2002 to FY 2008 inclusive.
5.
The tax effect should be included.
Propylene
[152]
Appellant produced the following table with
regard to propylene.
Sasol
Propylene, price / cost mark ups
Period of
analysis
Tier 1
Tier 2
Average
FY2001 to
FY2008
7.1%
19.6%
12.1%
FY2002 to
FY2008
8.8%
22.3%
14.3%
Sasol
Polypropylene
[153]
In respect of polypropylene appellant
presented a price – costs mark up under two scenarios which
were as follows:
·
Poly-A, which is the price / cost
comparison where the price of purified propylene is taken as the
value actually paid by Sasol
Polypropylene  (i.e. purified
propylene is purchased at the amount that is contained in the
financial statements); and
·
Poly-D2 (i.e. the integrated approach),
which is the price / costs comparison where the price of propylene is
set equal to the estimated
economic costs of propylene, assuming
feedstock propylene is set equal to the actual costs of feedstock
propylene as contained
in the financial statements of Sasol
Propylene.
[154]
Appellant’s calculation was set out
in the following table:
Sasol
Propylene, price / cost mark ups
Period of
analysis
Poly-A
Poly D
FY2001 to
FY2008
-28.9%
-28.2%
FY2002 to
FY2008
-27.2%
-24.4%
[155]
On 5 May 2015 this court received the
following letter from the Commission’s attorneys:
‘
We
refer to the Court’s letters of 17, 20 and 23 April 2015 and
SCI’s response of 24 April 2015 and the memorandum from
FTI
attached to that response.
The
Commission agrees with the figures contained in the FRI memorandum
(although the Commission does not  understand the
Court to
have requested the calculation of “Poly A”, which is not
on an integrated basis) and wishes only to draw to
the Court’s
attention two further assumptions made by FTI in the calculation.
Firstly,
as FTI records at the end of paragraph 1.3 of its memorandum, the
prices FTI has used in its calculation of the polypropylene

price/cost markup include sales made under the Customer Export
Incentive Programme (CEIP), which are at lower prices than domestic

sales, and which the Commission contends ought to be have been
excluded from the calculation (as set out in paragraphs 241 - 247
of
the Commission’s heads of argument).  If these sales are
excluded then the markups for polypropylene contained in
the FTI
memorandum should be increased by approximately 2%.
Secondly,
FTI also records at the end of paragraphs 1.3 of its memorandum that
it has used a weighted average in its calculations.
By this it
means that it has averaged prices across the years and also the costs
across the years before calculating the markup,
rather than giving an
equal weight to the percentage markup in each year as the Commission
contends ought to be done (as set out
in paragraphs 220 – 225
of the Commission’s heads of arguments).  Using the
Commission’s method increases
the markups reflected in the FTI
memorandum by less than 1% though and the difference in the two
approaches is therefore not material
to the calculations requested by
the Court.
[156]
The Commission also noted that:
‘
It
observed that on the basis of that the impact would be different if a
different permutation of the disputed amendments
was accepted
by the Court.  Mr Harman’s slide presentation the effect
of using appellants approach is to decrease the
average mark-up by
between 1.9% and 2.4%.’
[157]
Mention was made in the letter from the
Commission’s attorneys of the Customer Export Incentive Program
(CEIP).   This
now requires some attention.  Appellant
offered a rebate to customers on the price of PP if the latter
manufactured a product
using polypropylene for a product to be
exported.  The discount was offered on the basis that appellant
would otherwise be
obliged to export the PP itself.   The
Commission contended that it was not appropriate to include these
prices in the
calculation of prices that were the subject matter of
the complaint.  The complaint only concerned appellant’s
domestic
prices.  Furthermore, the Commission did not allege
that the CEIP prices, which are significantly lower than the domestic
prices, were excessive.  The Tribunal held that, if these prices
were excluded, the costs should be similarly excluded, which
the
Commission had not done.
[158]
Leaving aside the merit of the arguments
and assuming that these sales were excluded, the mark up for PP
should be increased by
approximately 2%.  Assuming further that
the Commission was correct that appellant should not have taken
average prices across
the years together with the costs before
calculating the mark up, but rather given equal weight to the
percentage mark up in each
year, as the Commission contended ought to
be done, this would have increased the mark up by less than 1%.
Furthermore,
on the assumption of a different permutation of
the amendments, as urged upon this Court by the Commission, this
would have increased
the average mark up by between 1.9 to 2.4%.
[159]
On the basis of all of these assumptions,
it would mean that the price-costs mark up on average would be in the
region of 16% for
FY 2001 to FY 2008 insofar as propylene is
concerned and from FY 2002 to FY 2008 to slightly less than 19%.
Commensurate
increases would also have to be made insofar as the
separate Tier 1 and Tier 2 calculations are concerned.
Manifestly, on
these calculations the changes to the PP price-cost
mark up should not vex any court with respect to an excessive pricing
dispute.
[160]
As this court has taken the average of Tier
1 and Tier 2 prices, for the reasons already advanced in this
judgment, on the basis
of the calculations provided by both of the
parties, irrespective of whether one takes FY 2001 to FY 2008 or FY
2002 to FY 2008
the price-cost mark up would be with the range of
between 12% - 14%.
Reasonableness
[161]
Given this finding, it is possible to turn
to the second leg of the enquiry, namely the question of
reasonableness; that is whether
appellant’s prices are
reasonably related to economic value.
[162]
As noted, s 8 (a) of the Act is clear: a
price is excessive only if it is higher than economic value and bears
no reasonable relation
thereto.   Section 8 (a) does not
apply to prohibit any price in excess of economic value.  This
conclusion must
follow from the wording that is employed:  the
price is excessive if it is both higher than economic value and bears
no reasonable
relation thereto.  This formulation, as has been
noted on a number of occasions in this judgment, was derived from the
decision
of the European Court of Justice in
United
Brands Company v EC Commission, supra
at para 248-250.  The relevant article of the EU Treaty, with
which excessive pricing is concerned is Article 102.  In
the
German version this article contains the following wording
‘Unangenessene Preisen’; that is prices that are out
of
line or disproportionate.  Accordingly, a proportionality
enquiry appears to be indicated to determine the reasonableness

component of the test for excessive pricing.
[163]
In its decision, the Tribunal was cognisant
of an observation of this Court in
Mittal
that a reasonable assessment involved a value judgment and that there
was no single inflexibly clear threshold which could be applied
to
determine whether a price was excessive in each and every case.
[164]
The Tribunal in its determination found
that the following features of the relevant market were important:
1. Appellant was a
dominant firm for the manufacture and supply of purified propylene.
It was also a dominant firm in the
production of sale of PP in South
Africa.
2. The barriers to
entry in these markets are high.  Access to purified propylene
and polypropylene markets would require reliable
and significant
supply feedstock propylene.  There was no supply in South
African other than by way of Synfuels.
3. Appellant enjoyed
a cost advantage as the producer of purified propylene and PP due to
the nature of the feedstock propylene
as a by-product of Synfuels’
fuel operations.
4. Synfuels had an
abundance of feedstock propylene as a bi product and poor alternative
uses therefore
5. Appellant had not
been able to demonstrate that its market position in purified
propylene and PP with the result of innovation
or risk taking on its
part.
6. Sasol had enjoyed
very significant state support for its business for a protracted
period of time.  It had leveraged this
support to create
positions of dominance in the domestic market for both purified
propylene and PP.
7. High input prices
for both purified propylene and PP held significant implications for
the South African economy since they had
marked negative effects on
the relevant downstream industries.
[165]
The Tribunal then held: ‘In the case
of purified propylene the price-cost test is the only reliable
indicator of the economic
value of the purified propylene sold by SCI
during the complaint period.’   As it had found that
the final price
cost test for purified propylene showed that prices
over actual costs were in the range of between 39.9% to 51.5% for
Tier 2 sale
to Safripol and in the range of 25.1 to 26.5 for Tier 1
sales to Safripol, the Tribunal was satisfied that this was both
excessive
and unreasonable in relation to economic value. With regard
to PP it found that during the complaint period the price over actual

costs was in the range of between 17.6% and 25.4 %, on a very
conservative basis and between 26.9% to 36.5% on a more realistic

basis.  Furthermore, a comparison of appellant’s domestic
polypropylene prices to the prices in Western European indicated
that
appellant’s domestic prices were 41% and 47% higher
respectively for the homopolymer and raffia grade in the relevant

period compared to the Western European discounted prices computed on
the basis of feedstock cost comparable to appellant.
On this
basis it found that there was no reasonable relationship between the
price charged by appellant to the local plastic convertors
for PP and
the economic value thereof.
[166]
Certain conclusions with respect to this
leg of the overall inquiry which were reached by the Tribunal
unfortunately do not take
the matter particularly far.  The fact
that appellant was dominant will always be the case when dealing with
an excessive
pricing dispute in terms of s 8 (a). To the argument
that propylene and PP are characterised by high barriers to the
entry, in
particular because entry requires access to a significant
supply of feedstock propylene, Dr Padilla showed that there was in
fact
an excess of feedstock propylene supply in the market during the
complaint period.  The Tribunal appeared to have agreed that

there was an abundance of feedstock propylene in the market.
[167]
Mr Trengove also observed correctly that
appellant sold propylene to Safripol and there was no reason to think
that it would not
have given serious consideration to supplying a new
entry into the market, as might have been the case with Project
Mango.
Further, with regard to the question as to whether the
prices caused harm to consumers, this was itself a separate enquiry
to that
dealing with reasonableness as, again, is evident from the
wording of s 8 (a).
[168]
Throughout his argument Mr Wesley pressed
the point concerning the difference between the prices charged by
international producers
as compared to appellant.  In
particular, Mr Wesley referred to the fact that the net prices for PP
in Europe were between
12% to 18% lower than appellant’s
domestic prices.
[169]
Mr Wesley pointed out that Dr
Padilla’s calculation in this regard included the CEIP rebate,
which, in his view, should not
have been included.   If
this rebate was excluded, appellant’s prices were between 15.4
and 20.5% higher than Western
European prices in the complaint period
for homopolymer and raffia grade respectively.  Mr Wesley noted
that it was not in
dispute that European firms were not low cost
producers.
[170]
This argument is one which, as this court
pointed out in
Mittal,
serves as a measure to test economic value.   See para 51
of the
Mittal
judgment.  Nonetheless, this would not prevent the Commission
from employing this comparator to justify its conclusion with
respect
to the finding that there was no reasonable relationship between the
price charged by appellants to the local plastic convertors
for PP
during the complaint period and the economic value of the PP.
But as already noted, appellant’s price-cost
figures are not in
favour of the Commissions’ case with respect to economic value.
[171]
A great deal was made by the Commission and
by the Tribunal with regard to the origins of appellant’s
dominance, in particular
the history of state support and the fact
that appellant’s dominant position in the relevant market was
not the result of
any innovation or risk taking on its part.
This Court in
Mittal
considered that these factors should be examined at the
reasonableness stage of the enquiry, because it was here that it was
appropriate
to take into account how the firm’s cost affected
the reasonableness of its price in relation to the value of the good
and
whether the high price of the good represented a reward for risk
and innovation.
[172]
Unfortunately the Tribunal’s emphasis
on innovation and risk taking resulted in it in making an observation
which is manifestly
incorrect, for if followed, it would destroy any
possibility of instituting a successful excessive pricing case in the
important
area of intellectual property.  This observation is
the following:
‘
Due
regard must be paid to the fact that a dominant firm’s price
for the product or service may justifiably be higher than
its
economic value.  An example would be the pricing of a patented
product where the patent holder has the right to the economic

exploitation of the innovation for a limited period.
Accordingly, a patent holder may charge a price which bears no
relation
to the economic value of the product for the duration of the
specific patent.  This, however, is not a relevant factor here.
’
[173]
As Sutherland and Kemp
Competition
Law of South Africa
(loose-leaf at 7-50
(4)) correctly observe:
‘
While
patent holder innovation research expenditure may have a bearing on
economic value of its product and the reasonableness of
its price
this is not, as the Tribunal seems to suggest, a license for patent
holders to engage in excessive pricing.  Such
an approach has no
basis on the wording of s 8 (a).
’
[174]
In this case however, even if this court
was inclined to consider that a price costs ratio of 14% is
excessive, this on its own
cannot suffice to justify a holding that
the price bears no reasonable relation to the economic value of the
good or service.
Of course, it must be remembered that a
return of 14% is in addition to the return on capital which has been
factored into the
costs and hence into the calculation of economic
value.  Thus, the figure of 14% is in addition to a competitive
return which
is the reason why it must be regarded as excessive in
relation to economic value.   But returns above economic
value
are not
per se
unreasonable.  In this case, when appellant manufactured
purified propylene in its Monomers division, it sold the propylene
to
its Polymers division and to Safripol at the same prices.  Its
PP prices were only approximately 16% above the cheapest
PP in the
world.  Half of this difference was as a result of government
policy which maintained a tariff protection on imported
PP throughout
the complaint period, a fact which was confirmed by Dr Roberts under
cross examination.
[175]
A review of the European jurisprudence
indicates that the prices charged have to be substantially higher
than the defined economic
value before an adverse finding will be
made.  See for example,
British
Leyland Plc. v Commission
1987 (1) CMCR
185;
Napp Pharmaceutical Holdings Ltd v
Director of Fair Trading
[2002] EWCA Civ 796
;
[2002] 4 All
ER 376.
A price which is significantly less than 20% of the
figure employed to determine economic value falls short of justifying

judicial interference in this complex area.  Ancient support can
be found for this finding.  The TALMUD (Baba Bathra
90a) ruled
that if the profit gained was more than 16.67% it was regarded as
excessive.
[176]
For all of these reasons therefore the
evidence as it is presented to the Tribunal cannot justify the
finding that the price of
propylene bore no reasonable relation to
the economic value of the good.
[177]
On the basis of this finding, it is
unnecessary to engage in the question as to whether there was
detriment to consumers nor is
the Court required to deal with the
question of administrative penalties.
Economic
experts
[178]
Throughout this judgment I have raised
questions of expert evidence and specifically the role of an expert
in these proceedings.
This important question is not only
a feature of this case but, unfortunately, has raised itself in
numerous appeal records which
have been presented to this Court over
the past fifteen years.    Regrettably this Court must
now draw attention
to the inability of the Tribunal to exercise
discipline over proceedings in order to in order to ensure that
economic experts provide
evidence on economic questions, leaving
points of legal interpretation to the Tribunal and to this Court,
after submissions by
the extremely qualified lawyers who represent
the parties have been presented.
[179]
There are further issues with regard
to the expert evidence which is led before the Tribunal.  For
the guidance of experts
who present evidence before the Tribunal in
future cases as well as the Tribunal itself, I propose to deal with
this question.
I readily accept that there are areas of
economics where there is an overwhelming consensus over the
uncompetitive character of
certain business practices in a particular
context.  But as Richard Posner “
The
Law and Economics of the Economic Expert Witness
”
1999 (13)
The Journal of Economic
Perspectives
91 at 96 has written:
‘
Where
the use of economic experts is more problematic is in the areas of
economics where there is no professional consensus.
This used
to be and to some extent still is the situation with regard to
antitrust economics. A perfectly  respected economist
may be an
antitrust “hawk”, another equally respected economic
economist an antitrust dove.  Each might have a
long list of
reputable academic publications fully consistent with systematically
pro-plaintiff or pro-defendant testimony, and
so a judge or jury
would have little basis for choosing between them.’
Closer
to the reasons which have prompted this excursus, Barbier de la Serre
and Sibony have observed that there are numerous cases
‘in
which the conclusions of the experts reports were not irrelevant but
were questioned  and/or judge unfounded (eg
when the other party
submitted an expert report that contradicted the findings of the
other report, the report did not put forwards
the ‘slightest
evidence’ supporting its conclusion, the experts’
conclusions were based on complex premises which
in view of their
number and  complexity did not permit sufficiently defined
conclusions, the experts qualifications did not
correspond to the
factual issues at stake and the report was based on incomplete
knowledge where the facts were simply ‘unreliable’.

(‘Expert evidence before the EC Courts’
(2008) 45
Common
Market Law Review
941
at 968;  See
also
Impala v Commission
[2006] ECR II – 2289 at para 345.
[180]
This concern about the impartiality of
expertise and its implications for complex cases such as the present
were captured by Lord
Woolf in his
Interim
Access to Justice Report (Interim Report to the Lord Chancellor on
the Civil Justice System) in England and in Wales
(1995).
See Chapter 23 at para 5:
‘
Most
of the problems with expert evidence arise because the expert is
initially recruited as part of the team which investigates
and
advances a party’s contentions and then has to change roles and
seek to provide the independent expert evidence
which the court
is entitled to expect.  As Lord Wilberforce in
The
Ikarian Reefer
(1993) 2 Lloyds Reports
68) stated:  ‘It is necessary that expert evidence
presented to the court should be and should
be seen to be the
independent product of the expert uninfluenced as to form or content
by the exigencies of litigation.’
In many cases the expert,
instead of playing the role identified by Lord Wilberforce has become
… ‘a very effective
weapon in the party’s arsenal
of tactics.’
[181]
This approach finds support in the leading
South African authority, Zeffertt and Paizes
The
South African Law of Evidence
(2
nd
ed) at 330 who cite favourably from
The
Ikarian Reefer
, which judgment they
consider reflects adequately the duties and responsibilities of
expert witnesses:
‘
1.
Expert evidence presented to the Court should be, and should be seen
to be, the independent product of the expert uninfluenced
as to form
or content by the exigencies of litigation (
Whitehouse
v Jordan
[1981] W.I.R. 246 at p.  256.
per Lord Wilberforce).
2.
An expert witness should provide independent assistance to the court
by way of objective unbiased opinion in relation to matters
within
his expertise (see
Polivitte Ltd v Commercial Union Assurance Co.
Plc
., [1987]  Lloyd’s  Rep. 379 at p. 386 per Mr
Justice Garland and
Re J.
[1990] F.C.R. 193
per Mr Justice
Cazalet).  An expert witness in the High Court should never
assume the role of an advocate.
3.
An expert witness should state the facts or assumption upon which his
opinion is based.  He should not omit to consider
material facts
which could detract from his concluded opinion (
Re J
sup.).
4.
An expert witness should make it clear when a particular quotation or
issue  falls outside his expertise.
5.
If an expert’s opinion is not properly researched because he
considers that insufficient data is available, then this must
be
stated with an indication that the opinion is no more than a
provisional one.  (
Re J
sup.),  In cases where an expert witness who  has prepared
a report could not assert that the report contained the truth,
the
whole truth and nothing but the truth without some qualification,
that qualification should be stated in the report (
Derby
& Co. Ltd. and Others v Weldon and Others
,
The Times
,
Nov. 9, 1990 per Lord Justice Staughton)…’
[182]
It is these guidelines which should be
followed in future hearings before the Tribunal.  Suffice to say
that in this case,
once the issue of feedstock costs had been
resolved, the appellant placed before the Tribunal impressive and
cogent evidence which,
sadly was not properly gainsaid by the
Commission, to the extent that it would have been justified to
conclude, on the probabilities,
that the Commission’s case
should have been upheld.   To the contrary, figures cited
without any clear and reasoned
justification do not constitute expert
evidence.  Further, an expert in a defined area is not
necessarily an expert in another
defined area.  The Tribunal
must in future guard against expert overreaching.
[183]
These observations are made partly to
counter the predictable reaction to a decision of this Court in this
complex area of excessive
pricing, namely that excessive cases can
never succeed before South African courts.  This would be a
completely unjustified
conclusion.  Had this court been
confronted with evidence to properly contest Mr Harman’s
testimony in particular and
thus had the Court enjoyed the benefit of
expert evidence, as set out above, with regard to the evaluation of
capital assets, the
level of the capital reward / return on capital,
the allocation of group and common cost, this evidence may this well
have shed
a different light on this case.
[184]
As indicated earlier, some measure of
latitude has to be given to firms with regard to pricing.  If
not, a court will become
a price regulator; hence the importance of
sustained expert evidence.  Furthermore, to the extent that the
policy issues are
important, particularly within the context of the
South African economy the observation made by Sutherland et al should
also be
considered:
‘
The
Tribunal distinguished between firms that have gained their dominance
through state support and those that have achieved dominance
through
innovation and risk-taking.  Given the passage of time and
firm-specific developments since the receipt of such support,
it may
be difficult for a dominant firm accurately to determine the source
of its dominance, or the extent to which any subsequent
innovation
and risk-taking may weigh against that support.  And yet, as
outlined above, this distinction may fundamentally
alter the economic
value of the relevant good and thus the firm’s obligations
under the Act. An appeal has been lodged in
the
Sasol
case.’
(at 7 – 50 (4))
Safripol
[185]
The second appellant, Safripol noted an
appeal against the behavioural remedies granted by the Tribunal in
paragraphs 455 to 509
of its decision.   Safripol seeks an
order setting aside the behavioural remedies granted by the Tribunal
and replacing
them with a set of behavioural remedies which it
proposes to the Court.  Safripol does not however appeal against
the remainder
of the Tribunal’s decision.  Indeed,
Safripol supported the Tribunal’s finding that first appellant
has engaged
in excessive pricing.   Given the approach
adopt by this Court to the appeal of appellant, there is no need to
deal with
the arguments put forward by Safripol regarding the
Tribunal’s order.
Conclusion
[186]
In the result, the appeal is upheld. The
decision of the Competition Tribunal is set aside and replaced with
the following order:
The
complaint referral is dismissed
DAVIS
JP
MOLEMELA
and VICTOR AJJA concurred